------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  /Users/rommel/Downloads/PRW_2020_ISQ_Replication/PRW_2020_ISQ_log.log
  log type:  text
 opened on:  29 Aug 2020, 12:59:27

. do "/var/folders/v5/jjp4w5d52p55gj12m0667jkc0000gn/T//SD02917.000000"

. *------------------------------------------------------------------------------*
. *       Paper:                  International Trade and Public Protest
. *                                       Evidence from Russian Regions
. *       Authors:                Tabea Palmtag, Tobias Rommel, Stefanie Walter
. *       Version:                June 10, 2020
. *       Task:                   Empirical Analysis
. *------------------------------------------------------------------------------*
. 
. 
. clear all

. version 14

. set mat 3000

. set more off

. 
. 
. 
. 
. ***** 1. Dataset
. *------------------------------------------------------------------------------*
. 
. *** Set directory
. local dir = "/Users//`=c(username)'/Downloads/PRW_2020_ISQ_Replication/"

. cd "`dir'"
/Users/rommel/Downloads/PRW_2020_ISQ_Replication

. 
. *** Load data
. use "PRW_2020_ISQ_regiondata.dta", clear
(Palmtag/Rommel/Walter 2020)

. 
. *** Graph preferences
. * ssc install blindschemes
. * set scheme plotplain
. 
. 
. 
. ***** 2. Data Preparation
. *------------------------------------------------------------------------------*
. 
. *** Convert regional GDP from RUB to USD
. gen grp=reg_grp*(xr_rus)^(-1)
(761 missing values generated)

. lab var grp "Gross regional product (in million USD)"

. 
. *** Trade openness
. * Imports and exports
. gen tradeopen=((reg_exporttofor+reg_exporttosng+ ///
>         reg_importtofor+reg_importtosng)/grp)*100
(883 missing values generated)

. lab var tradeopen "Imports+Exports (% of GDP)"

. gen lntrade = ln(tradeopen + 1)
(883 missing values generated)

. lab var lntrade "Imports+Exports (% of GDP), ln"

. * Exports
. gen exportopen=((reg_exporttofor+reg_exporttosng)/grp)*100
(869 missing values generated)

. lab var exportopen "Exports (% of GDP)"

. gen lnexport = ln(exportopen + 1)
(869 missing values generated)

. lab var lnexport "Exports (% of GDP), ln"

. * Imports
. gen importopen=((reg_importtofor+reg_importtosng)/grp)*100
(868 missing values generated)

. lab var importopen "Imports (% of GDP)"

. gen lnimport = ln(importopen + 1)
(868 missing values generated)

. lab var lnimport "Imports (% of GDP), ln"

. 
. *** FDI openness
. gen fdiopen=((reg_fdi_incl_total/1000)/grp)*100
(1,326 missing values generated)

. lab var fdiopen "FDI inflows (% of GDP)"

. gen lnfdi = ln(fdiopen + 1)
(1,326 missing values generated)

. lab var lnfdi "FDI inflows (% of GDP), ln"

. 
. *** Generate average education level (secondary + tertiary)
. * Secondary and tertiary (main models)
. bysort rcode: egen edshare=mean(higheduc) if year>=2007
(1435 missing values generated)

. gen educshare = edshare-34
(1,435 missing values generated)

. * Tertiary (main models)
. bysort rcode: egen edshareter=mean(tereduc) if year>=2007
(1435 missing values generated)

. gen educshareter = edshareter-17
(1,435 missing values generated)

. * Secondary + 2*tertiary (main models)
. bysort rcode: egen edshareind=mean(higheduc+tereduc) if year>=2007
(1435 missing values generated)

. gen educshareind = edshareind-54
(1,435 missing values generated)

. * Secondary and tertiary (robustness checks)
. bysort rcode: egen edshare2=mean(higheduc) if year>=2003
(1115 missing values generated)

. gen educshare2 = edshare2-36
(1,115 missing values generated)

. * Secondary and tertiary (regional mechanism)
. bysort rcode: egen edshare4=mean(higheduc) if year>=2000
(875 missing values generated)

. gen educshare4 = edshare4-36
(875 missing values generated)

. 
. *** Generate GRP growth
. sort rcode year

. by rcode: gen reg_grpgr=(grp-grp[_n-1])/grp[_n-1]
(845 missing values generated)

. 
. *** Population size in millions
. replace reg_pop =reg_pop/1000000
variable reg_pop was long now double
(1,327 real changes made)

. 
. *** Population density
. gen reg_density = reg_pop/reg_area
(748 missing values generated)

. 
. *** Logarithm of GRP per capita
. gen lngrppc = ln(reg_grp_pc)
(793 missing values generated)

. 
. *** Newspaper coverage
. gen lnnews = ln(reg_newspaper+1)
(761 missing values generated)

. 
. *** Road density
. gen lnroadden = ln(reg_autoroadden)
(747 missing values generated)

. 
. *** Distance to Moscow
. gen lndistmos = ln(reg_disttom+1)
(747 missing values generated)

. 
. 
. 
. ***** 3. Descriptive Statistics
. *------------------------------------------------------------------------------*
. 
. *** Sample size
. xtset rcode year
       panel variable:  rcode (strongly balanced)
        time variable:  year, 1990 to 2014
                delta:  1 unit

. xtnbreg protest_ikd2 cl.lntrade##c.educshare l.lnfdi l.reg_pop ///
>         l.reg_urbanshare l.lngrppc l.reg_grpgr l.reg_levelofunempl ///
>         l.lnnews l.lndistmos l.lnroadden, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -34295.114  
Iteration 1:   log likelihood = -17862.205  (backed up)
Iteration 2:   log likelihood = -7609.2442  
Iteration 3:   log likelihood = -4440.2223  
Iteration 4:   log likelihood = -2818.9144  
Iteration 5:   log likelihood = -2804.8261  
Iteration 6:   log likelihood = -2804.7696  
Iteration 7:   log likelihood = -2804.7696  

Iteration 0:   log likelihood = -3860.5381  
Iteration 1:   log likelihood =   -1735.91  
Iteration 2:   log likelihood =  -1463.678  
Iteration 3:   log likelihood = -1462.4843  
Iteration 4:   log likelihood = -1462.4833  
Iteration 5:   log likelihood = -1462.4833  

Iteration 0:   log likelihood = -1462.4833  (not concave)
Iteration 1:   log likelihood = -1404.9366  (not concave)
Iteration 2:   log likelihood = -1377.0615  
Iteration 3:   log likelihood = -1328.8497  
Iteration 4:   log likelihood = -1324.3128  
Iteration 5:   log likelihood = -1324.2707  
Iteration 6:   log likelihood = -1324.2707  

Fitting full model:

Iteration 0:   log likelihood = -1444.0961  (not concave)
Iteration 1:   log likelihood = -1347.0788  
Iteration 2:   log likelihood =  -1261.626  
Iteration 3:   log likelihood = -1251.0022  
Iteration 4:   log likelihood = -1250.8217  
Iteration 5:   log likelihood = -1250.8216  
Iteration 6:   log likelihood = -1250.8216  

Random-effects negative binomial regression     Number of obs     =        428
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(12)     =      75.29
Log likelihood  = -1250.8216                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5139122   .2391714     2.15   0.032     .0451448    .9826796
                       |
             educshare |   .0682669   .0557621     1.22   0.221    -.0410247    .1775586
                       |
cL.lntrade#c.educshare |  -.0460403   .0172228    -2.67   0.008    -.0797963   -.0122843
                       |
                 lnfdi |
                   L1. |   .1184438   .0963921     1.23   0.219    -.0704812    .3073689
                       |
               reg_pop |
                   L1. |   .4660285   .0714359     6.52   0.000     .3260167    .6060403
                       |
        reg_urbanshare |
                   L1. |   .0179866   .0103543     1.74   0.082    -.0023074    .0382806
                       |
               lngrppc |
                   L1. |  -.6218817   .1267053    -4.91   0.000    -.8702194    -.373544
                       |
             reg_grpgr |
                   L1. |   -.457377   .2056261    -2.22   0.026    -.8603968   -.0543572
                       |
     reg_levelofunempl |
                   L1. |  -.0316325   .0290475    -1.09   0.276    -.0885647    .0252996
                       |
                lnnews |
                   L1. |   -.034887   .1111976    -0.31   0.754    -.2528303    .1830563
                       |
             lndistmos |
                   L1. |   .2154477   .0755081     2.85   0.004     .0674546    .3634408
                       |
             lnroadden |
                   L1. |  -.1239417   .0996336    -1.24   0.214      -.31922    .0713366
                       |
                 _cons |   4.715761   2.127807     2.22   0.027     .5453362    8.886186
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3682047   .1890806                     -.0023864    .7387959
                 /ln_s |   1.195193   .2672814                      .6713312    1.719055
-----------------------+----------------------------------------------------------------
                     r |   1.445138   .2732475                      .9976164    2.093413
                     s |   3.304196   .8831502                       1.95684    5.579254
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 146.90               Prob >= chibar2 = 0.000

. gen sample = e(sample)==1

. 
. *** Table A1 in the Online Appendix
. tabstat protest_ikd2 protest_ikd_econ protest_mma protest_icews protest_kprf2 ///
>         lntrade educshare lnfdi reg_pop reg_urbanshare lngrppc reg_grpgr ///
>         reg_levelofunempl lnnews lndistmos lnroadden rents pressfill ///
>         KPRFmandateshare pctRussian if sample==1, statistics(N mean sd min max) ///
>         format(%9.2f)

   stats |  prote~d2  protes~n  protes~a  protes~s  prote~f2   lntrade  educsh~e     lnfdi
---------+--------------------------------------------------------------------------------
       N |    428.00    428.00    428.00    428.00    428.00    422.00    428.00    424.00
    mean |     14.11      5.34      3.67      3.36     10.76      3.19     12.91      0.66
      sd |     41.58     16.21     11.47     10.56      9.74      0.78      5.75      0.58
     min |      0.00      0.00      0.00      0.00      0.00      0.65      0.83      0.00
     max |    406.00    173.00    112.00    111.00     62.00      5.10     29.93      3.55
------------------------------------------------------------------------------------------

   stats |   reg_pop  reg_ur~e   lngrppc  reg_gr~r  reg_le~l    lnnews  lndist~s  lnroad~n
---------+--------------------------------------------------------------------------------
       N |    427.00    428.00    428.00    427.00    428.00    428.00    428.00    428.00
    mean |      1.93     71.43     12.16      0.14      7.07      6.63      6.97      4.50
      sd |      1.75     10.93      0.53      0.20      2.55      0.73      1.61      1.32
     min |      0.05     27.10     11.10     -0.43      0.80      4.58      0.00     -0.22
     max |     11.92    100.00     14.10      0.83     21.70      8.94      9.38      7.68
------------------------------------------------------------------------------------------

   stats |     rents  pressf~l  KPRFma~e  pctRus~n
---------+----------------------------------------
       N |    428.00    426.00    419.00    428.00
    mean |      7.37      2.05     10.15     82.93
      sd |     12.00      0.66      6.07     16.65
     min |      0.00      1.00      0.00      9.20
     max |     55.70      3.00     33.00     90.00
--------------------------------------------------

. 
. *** Correlation between protest indicators
. corr protest_ikd2 protest_ikd_econ protest_mma protest_icews protest_kprf2 ///
>         if sample==1
(obs=428)

             | prote~d2 protes~n protes~a protes~s prote~f2
-------------+---------------------------------------------
protest_ikd2 |   1.0000
protest_ik~n |   0.9544   1.0000
 protest_mma |   0.7838   0.7076   1.0000
protest_ic~s |   0.4858   0.3703   0.7704   1.0000
protest_k~f2 |   0.5010   0.4708   0.5064   0.3893   1.0000


. 
. *** Histograms
. hist protest_ikd2 if sample==1, ///
>         xtitle("Number of grassroots IKD protests") ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(hist_ikd, replace) fysize(40)
(bin=20, start=0, width=20.3)

. hist tradeopen if sample==1, ///
>         xtitle("Exposure to international trade") ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(hist_trade, replace) fysize(40)
(bin=20, start=.90795887, width=8.0898712)

. 
. *** Prepare trade and FDI exposure
. preserve

. * Keep protest and trade variables
. keep rname rid year protest_ikd2 protest_kprf2 tradeopen lntrade fdiopen lnfdi

. keep if year>=2000 & year!=2014
(913 observations deleted)

. * Generate merge variable for individual-level data
. gen conversion =        .               
(1,162 missing values generated)

. replace conversion =    1       if rname ==     "Altai K"
(14 real changes made)

. replace conversion =    2       if rname ==     "Amur"
(14 real changes made)

. replace conversion =    3       if rname ==     "Chelyabinsk"
(14 real changes made)

. replace conversion =    4       if rname ==     "Chuvash"
(14 real changes made)

. replace conversion =    5       if rname ==     "Kabardino-Balkar"
(14 real changes made)

. replace conversion =    6       if rname ==     "Kaliningrad"
(14 real changes made)

. replace conversion =    7       if rname ==     "Kaluga"
(14 real changes made)

. replace conversion =    8       if rname ==     "Komi"
(14 real changes made)

. replace conversion =    9       if rname ==     "Krasnodar"
(14 real changes made)

. replace conversion =    10      if rname ==     "Krasnoyarsk"
(14 real changes made)

. replace conversion =    11      if rname ==     "Kurgan"
(14 real changes made)

. replace conversion =    12      if rname ==     "Leningrad"
(14 real changes made)

. replace conversion =    13      if rname ==     "Lipetsk"
(14 real changes made)

. replace conversion =    14      if rname ==     "Moscow C"
(14 real changes made)

. replace conversion =    15      if rname ==     "Moscow O"
(14 real changes made)

. replace conversion =    16      if rname ==     "Nizhny Novgorod"
(14 real changes made)

. replace conversion =    17      if rname ==     "Novosibirsk"
(14 real changes made)

. replace conversion =    18      if rname ==     "Orenburg"
(14 real changes made)

. replace conversion =    19      if rname ==     "Penza"
(14 real changes made)

. replace conversion =    20      if rname ==     "Perm"
(14 real changes made)

. replace conversion =    21      if rname ==     "Primorsky"
(14 real changes made)

. replace conversion =    22      if rname ==     "Rostov"
(14 real changes made)

. replace conversion =    23      if rname ==     "Saint Petersburg"
(14 real changes made)

. replace conversion =    24      if rname ==     "Saratov"
(14 real changes made)

. replace conversion =    25      if rname ==     "Smolensk"
(14 real changes made)

. replace conversion =    26      if rname ==     "Stavropol"
(14 real changes made)

. replace conversion =    27      if rname ==     "Tambov"
(14 real changes made)

. replace conversion =    28      if rname ==     "Tatarstan"
(14 real changes made)

. replace conversion =    29      if rname ==     "Tomsk"
(14 real changes made)

. replace conversion =    30      if rname ==     "Tula"
(14 real changes made)

. replace conversion =    31      if rname ==     "Tyumen"
(14 real changes made)

. replace conversion =    32      if rname ==     "Udmurt"
(14 real changes made)

. replace conversion =    33      if rname ==     "Volgograd"
(14 real changes made)

. save "glob_stats.dta", replace
file glob_stats.dta saved

. * Prepare data for maps
. keep if year>=2007 & year<=2012
(664 observations deleted)

. collapse (first) rname (sum) protest_ikd2 protest_kprf2 (mean) tradeopen ///
>         lntrade fdiopen lnfdi, by(rid)

. save "map_stats.dta", replace
file map_stats.dta saved

. 
. *** Maps
. * ssc install shp2dta
. * ssc install geo2xy
. * ssc install spmap
. * Load shapefile
. shp2dta using "PRW_2020_ISQ_RUSadm1.shp", ///
>         database("map_data") ///
>         coordinates("map_coord") ///
>         genid(id) replace
type: 5

. * Prepare coordinates file
. use "map_coord", clear

. replace _X = 180 if _X<0
(56,292 real changes made)

. geo2xy _Y _X if _X>=-180 & _X<=180, replace projection(albers)

. drop if _Y<=4000000
(5 observations deleted)

. save "map_coord", replace
file map_coord.dta saved

. * Prepare region identifiers
. use "map_data", clear

. gen rname = NAME_1

. replace rname =         "Adygea"        if rname ==     "Adygey"
(1 real change made)

. replace rname =         "Altai K"       if rname ==     "Altay"
(1 real change made)

. replace rname =         "Altai R"       if rname ==     "Gorno-Altay"
(1 real change made)

. replace rname =         "Arkhangelsk"   if rname ==     "Arkhangel'sk"
(1 real change made)

. replace rname =         "Buryatia"      if rname ==     "Buryat"
(1 real change made)

. replace rname =         "Chechen"       if rname ==     "Chechnya"
(1 real change made)

. replace rname =         "Chukotka"      if rname ==     "Chukot"
(1 real change made)

. replace rname =         "Ingushetia"    if rname ==     "Ingush"
(1 real change made)

. replace rname =         "Kabardino-Balkar"      if rname ==     "Kabardin-Balkar"
(1 real change made)

. replace rname =         "Kalmykia"      if rname ==     "Kalmyk"
(1 real change made)

. replace rname =         "Khakassia"     if rname ==     "Khakass"
(1 real change made)

. replace rname =         "Khanty-Mansi"  if rname ==     "Khanty-Mansiy"
(1 real change made)

. replace rname =         "Mari El"       if rname ==     "Mariy-El"
(1 real change made)

. replace rname =         "Magadan"       if rname ==     "Maga Buryatdan"
(1 real change made)

. replace rname =         "Moscow C"      if rname ==     "Moscow City"
(1 real change made)

. replace rname =         "Moscow O"      if rname ==     "Moskva"
(1 real change made)

. replace rname =         "Nizhny Novgorod"       if rname ==     "Nizhegorod"
(1 real change made)

. replace rname =         "North Ossetia-Alania"  if rname ==     "North Ossetia"
(1 real change made)

. replace rname =         "Oryol" if rname ==     "Orel"
(1 real change made)

. replace rname =         "Primorsky"     if rname ==     "Primor'ye"
(1 real change made)

. replace rname =         "Saint Petersburg"      if rname ==     "City of St. Petersburg"
(1 real change made)

. replace rname =         "Ulyanovsk"     if rname ==     "Ul'yanovsk"
(1 real change made)

. replace rname =         "Yamalo-Nenets" if rname ==     "Yamal-Nenets"
(1 real change made)

. replace rname =         "Zabaykalsky"   if rname ==     "Zabaykal'ye"
(1 real change made)

. replace rname =         "Astrakhan"     if rname ==     "Astrakhan'"
(1 real change made)

. replace rname =         "Perm"  if rname ==     "Perm'"
(1 real change made)

. replace rname =         "Ryazan"        if rname ==     "Ryazan'"
(1 real change made)

. replace rname =         "Stavropol"     if rname ==     "Stavropol'"
(1 real change made)

. replace rname =         "Tver"  if rname ==     "Tver'"
(1 real change made)

. replace rname =         "Tyumen"        if rname ==     "Tyumen'"
(1 real change made)

. replace rname =         "Yaroslavl"     if rname ==     "Yaroslavl'"
(1 real change made)

. replace rname =         "Jewish Autonomous"     if rname ==     "Yevrey"
(1 real change made)

. drop if rname==         "Sverdlovsk" & CCA_1!="0"
(2 observations deleted)

. save "map_data", replace
file map_data.dta saved

. * Merge descriptives
. use "map_stats", clear
(Palmtag/Rommel/Walter 2020)

. merge 1:1 rname using "map_data"

    Result                           # of obs.
    -----------------------------------------
    not matched                             0
    matched                                83  (_merge==3)
    -----------------------------------------

. drop _merge

. spmap protest_ikd2 using "map_coord" if rname!="Kaliningrad", ///
>         id(id) clnumber(9) fcolor(Greys2) ///
>         ocolor(white white white white white white white white white) ///
>         osize(vthin vthin vthin vthin vthin vthin vthin vthin vthin) ///
>         legenda(off) title("Grassroots IKD Protest") name(map_ikd, replace)

. spmap tradeopen using "map_coord" if rname!="Kaliningrad", ///
>         id(id) clnumber(9) fcolor(Greys2) ///
>         ocolor(white white white white white white white white white) ///
>         osize(vthin vthin vthin vthin vthin vthin vthin vthin vthin) ///
>         legenda(off) title("International Trade Exposure") name(map_trade, replace)

. 
. *** Figure 2 in the Main Text
. graph combine map_ikd map_trade hist_ikd hist_trade, col(2) ysize(3) xsize(5.5)

. graph save "figure_descriptives", replace
(note: file figure_descriptives.gph not found)
(file figure_descriptives.gph saved)

. restore

. 
. 
.         
. ***** 4. Protest models
. *------------------------------------------------------------------------------*
. 
. *** Model specification for IKD and KPRF
. * Dataset structure
. xtset rcode year
       panel variable:  rcode (strongly balanced)
        time variable:  year, 1990 to 2014
                delta:  1 unit

. * Baseline specifications
. global X1 cl.lntrade  c.educshare

. global X2 cl.lntrade##c.educshare

. global C1 l.reg_pop l.reg_urbanshare l.lngrppc l.reg_grpgr ///
>         l.reg_levelofunempl l.lnnews l.lndistmos l.lnroadden reg_density l.lnfdi 

. global C2 rents pressfill KPRFmandateshare pctRussian 

. * Sample size
. xtnbreg protest_ikd2 $X2 $C1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76824.046  (not concave)
Iteration 1:   log likelihood = -27716.233  
Iteration 2:   log likelihood = -12273.667  (backed up)
Iteration 3:   log likelihood = -8379.5418  
Iteration 4:   log likelihood = -2973.5931  
Iteration 5:   log likelihood = -2745.9373  
Iteration 6:   log likelihood = -2742.4212  
Iteration 7:   log likelihood = -2742.4186  
Iteration 8:   log likelihood = -2742.4186  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4794  (not concave)
Iteration 2:   log likelihood = -1397.7408  
Iteration 3:   log likelihood =  -1319.773  
Iteration 4:   log likelihood = -1315.7277  
Iteration 5:   log likelihood = -1315.6906  
Iteration 6:   log likelihood = -1315.6906  

Fitting full model:

Iteration 0:   log likelihood = -1415.3484  (not concave)
Iteration 1:   log likelihood = -1412.9143  (not concave)
Iteration 2:   log likelihood = -1298.8396  (not concave)
Iteration 3:   log likelihood = -1274.6434  
Iteration 4:   log likelihood = -1253.4921  
Iteration 5:   log likelihood =  -1249.492  
Iteration 6:   log likelihood = -1249.3982  
Iteration 7:   log likelihood = -1249.3981  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =      73.64
Log likelihood  = -1249.3981                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5241001   .2406021     2.18   0.029     .0525287    .9956715
                       |
             educshare |   .0719435   .0558374     1.29   0.198    -.0374958    .1813828
                       |
cL.lntrade#c.educshare |  -.0477285   .0174113    -2.74   0.006    -.0818539    -.013603
                       |
               reg_pop |
                   L1. |   .4604118   .0722552     6.37   0.000     .3187942    .6020294
                       |
        reg_urbanshare |
                   L1. |   .0167301   .0105232     1.59   0.112    -.0038949    .0373551
                       |
               lngrppc |
                   L1. |  -.6059637   .1288986    -4.70   0.000    -.8586003   -.3533272
                       |
             reg_grpgr |
                   L1. |  -.4506051   .2050263    -2.20   0.028    -.8524493   -.0487609
                       |
     reg_levelofunempl |
                   L1. |  -.0329513   .0290634    -1.13   0.257    -.0899145    .0240118
                       |
                lnnews |
                   L1. |  -.0436215   .1127103    -0.39   0.699    -.2645295    .1772865
                       |
             lndistmos |
                   L1. |   .2066014   .0763585     2.71   0.007     .0569414    .3562614
                       |
             lnroadden |
                   L1. |  -.1243937   .0995465    -1.25   0.211    -.3195013    .0707139
                       |
           reg_density |   .0432779   .0845487     0.51   0.609    -.1224345    .2089902
                       |
                 lnfdi |
                   L1. |   .1066998   .0977078     1.09   0.275    -.0848039    .2982034
                       |
                 _cons |   4.752546   2.128839     2.23   0.026     .5800993    8.924993
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3721707   .1911184                     -.0024146    .7467559
                 /ln_s |   1.190773   .2705402                      .6605236    1.721022
-----------------------+----------------------------------------------------------------
                     r |   1.450881     .27729                      .9975883    2.110143
                     s |   3.289622   .8899748                      1.935806    5.590236
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 132.59               Prob >= chibar2 = 0.000

. gen sample1 = e(sample)==1

. 
. *** Model specification for MMA and ICEWS
. * Model setup
. global X3 cl.lntrade  c.educshare2

. global X4 cl.lntrade##c.educshare2

. * Sample size
. xtnbreg protest_mma $X4 $C1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -24784.959  (not concave)
Iteration 1:   log likelihood = -8955.8889  
Iteration 2:   log likelihood =  -3677.478  (backed up)
Iteration 3:   log likelihood = -2102.0724  
Iteration 4:   log likelihood = -1329.8097  
Iteration 5:   log likelihood = -1215.4184  
Iteration 6:   log likelihood = -1214.9438  
Iteration 7:   log likelihood = -1214.9437  
Iteration 8:   log likelihood = -1214.9437  

Iteration 0:   log likelihood = -1836.2027  
Iteration 1:   log likelihood = -1400.4804  
Iteration 2:   log likelihood = -1285.9868  
Iteration 3:   log likelihood = -1225.5112  
Iteration 4:   log likelihood = -1225.4095  
Iteration 5:   log likelihood = -1225.4095  

Iteration 0:   log likelihood = -1225.4095  (not concave)
Iteration 1:   log likelihood = -1159.0426  
Iteration 2:   log likelihood = -1156.9629  
Iteration 3:   log likelihood = -1105.8907  
Iteration 4:   log likelihood = -1083.6773  
Iteration 5:   log likelihood = -1026.3963  
Iteration 6:   log likelihood = -1020.3206  
Iteration 7:   log likelihood =  -1020.268  
Iteration 8:   log likelihood =  -1020.268  

Fitting full model:

Iteration 0:   log likelihood = -1039.5493  
Iteration 1:   log likelihood = -1005.0949  
Iteration 2:   log likelihood = -999.34246  
Iteration 3:   log likelihood = -998.93423  
Iteration 4:   log likelihood = -998.93191  
Iteration 5:   log likelihood = -998.93191  

Random-effects negative binomial regression     Number of obs     =        561
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        7.5
                                                              max =          8

                                                Wald chi2(13)     =     203.64
Log likelihood  = -998.93191                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
            protest_mma |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                lntrade |
                    L1. |   .3240903   .2082733     1.56   0.120    -.0841178    .7322984
                        |
             educshare2 |   .1119427   .0559374     2.00   0.045     .0023075    .2215779
                        |
cL.lntrade#c.educshare2 |  -.0283557   .0174516    -1.62   0.104    -.0625603    .0058489
                        |
                reg_pop |
                    L1. |   .3337534   .0548093     6.09   0.000     .2263292    .4411776
                        |
         reg_urbanshare |
                    L1. |   .0349825   .0089664     3.90   0.000     .0174087    .0525563
                        |
                lngrppc |
                    L1. |   .4833978   .1024097     4.72   0.000     .2826784    .6841172
                        |
              reg_grpgr |
                    L1. |   .7775584   .2017475     3.85   0.000     .3821406    1.172976
                        |
      reg_levelofunempl |
                    L1. |   .1497319   .0271468     5.52   0.000     .0965252    .2029386
                        |
                 lnnews |
                    L1. |   .1420055   .0988956     1.44   0.151    -.0518262    .3358372
                        |
              lndistmos |
                    L1. |   .2607516   .0746388     3.49   0.000     .1144623     .407041
                        |
              lnroadden |
                    L1. |   .2457579   .0863576     2.85   0.004     .0765001    .4150156
                        |
            reg_density |   .0766181   .0315333     2.43   0.015      .014814    .1384222
                        |
                  lnfdi |
                    L1. |  -.1152952   .0943792    -1.22   0.222    -.3002749    .0696846
                        |
                  _cons |   -14.8802   1.899563    -7.83   0.000    -18.60327   -11.15713
------------------------+----------------------------------------------------------------
                  /ln_r |   1.786629   .2838804                      1.230233    2.343024
                  /ln_s |   1.882194   .3472093                      1.201677    2.562712
------------------------+----------------------------------------------------------------
                      r |   5.969296   1.694566                      3.422028    10.41268
                      s |   6.567901   2.280436                      3.325688    12.97095
-----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 42.67                Prob >= chibar2 = 0.000

. gen sample2 = e(sample)==1

. 
. *** Regression models
. * IKD protest, unconditional
. xtnbreg protest_ikd2 $X1 $C1 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood =  -76060.07  (not concave)
Iteration 1:   log likelihood = -27455.688  
Iteration 2:   log likelihood = -11896.857  (backed up)
Iteration 3:   log likelihood = -7765.2781  
Iteration 4:   log likelihood = -2999.4278  
Iteration 5:   log likelihood =   -2787.85  
Iteration 6:   log likelihood =   -2781.36  
Iteration 7:   log likelihood = -2781.3557  
Iteration 8:   log likelihood = -2781.3557  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4568  (not concave)
Iteration 2:   log likelihood = -1397.7621  
Iteration 3:   log likelihood = -1323.2789  
Iteration 4:   log likelihood = -1318.9195  
Iteration 5:   log likelihood = -1318.8657  
Iteration 6:   log likelihood = -1318.8657  

Fitting full model:

Iteration 0:   log likelihood = -1416.6377  (not concave)
Iteration 1:   log likelihood = -1332.9814  
Iteration 2:   log likelihood = -1259.8247  
Iteration 3:   log likelihood = -1253.0818  
Iteration 4:   log likelihood =  -1252.974  
Iteration 5:   log likelihood = -1252.9739  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(12)     =      63.25
Log likelihood  = -1252.9739                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------
     protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
          lntrade |
              L1. |  -.0615603   .1119505    -0.55   0.582    -.2809791    .1578586
                  |
        educshare |  -.0737809   .0175343    -4.21   0.000    -.1081475   -.0394143
                  |
          reg_pop |
              L1. |   .4231242   .0720845     5.87   0.000     .2818412    .5644073
                  |
   reg_urbanshare |
              L1. |    .008141   .0104614     0.78   0.436     -.012363    .0286449
                  |
          lngrppc |
              L1. |  -.6014235   .1290164    -4.66   0.000    -.8542911    -.348556
                  |
        reg_grpgr |
              L1. |  -.4580105   .2093077    -2.19   0.029     -.868246    -.047775
                  |
reg_levelofunempl |
              L1. |  -.0309349   .0292748    -1.06   0.291    -.0883124    .0264426
                  |
           lnnews |
              L1. |  -.1027022   .1129916    -0.91   0.363    -.3241617    .1187573
                  |
        lndistmos |
              L1. |   .2190132   .0788973     2.78   0.006     .0643774     .373649
                  |
        lnroadden |
              L1. |    -.11391   .1021078    -1.12   0.265    -.3140376    .0862176
                  |
      reg_density |  -.0145481   .1199033    -0.12   0.903    -.2495543    .2204581
                  |
            lnfdi |
              L1. |   .1139936   .0998534     1.14   0.254    -.0817155    .3097026
                  |
            _cons |   7.475422   1.938854     3.86   0.000     3.675339    11.27551
------------------+----------------------------------------------------------------
            /ln_r |   .2914258   .1866167                     -.0743362    .6571878
            /ln_s |   1.058627    .264286                      .5406355    1.576618
------------------+----------------------------------------------------------------
                r |   1.338334   .2497555                      .9283595    1.929359
                s |    2.88241   .7617805                      1.717098    4.838562
-----------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 131.78               Prob >= chibar2 = 0.000

. est store ikd1

. * IKD protest, conditional
. xtnbreg protest_ikd2 $X2 $C1 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76824.046  (not concave)
Iteration 1:   log likelihood = -27716.233  
Iteration 2:   log likelihood = -12273.667  (backed up)
Iteration 3:   log likelihood = -8379.5418  
Iteration 4:   log likelihood = -2973.5931  
Iteration 5:   log likelihood = -2745.9373  
Iteration 6:   log likelihood = -2742.4212  
Iteration 7:   log likelihood = -2742.4186  
Iteration 8:   log likelihood = -2742.4186  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4794  (not concave)
Iteration 2:   log likelihood = -1397.7408  
Iteration 3:   log likelihood =  -1319.773  
Iteration 4:   log likelihood = -1315.7277  
Iteration 5:   log likelihood = -1315.6906  
Iteration 6:   log likelihood = -1315.6906  

Fitting full model:

Iteration 0:   log likelihood = -1415.3484  (not concave)
Iteration 1:   log likelihood = -1412.9143  (not concave)
Iteration 2:   log likelihood = -1298.8396  (not concave)
Iteration 3:   log likelihood = -1274.6434  
Iteration 4:   log likelihood = -1253.4921  
Iteration 5:   log likelihood =  -1249.492  
Iteration 6:   log likelihood = -1249.3982  
Iteration 7:   log likelihood = -1249.3981  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =      73.64
Log likelihood  = -1249.3981                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5241001   .2406021     2.18   0.029     .0525287    .9956715
                       |
             educshare |   .0719435   .0558374     1.29   0.198    -.0374958    .1813828
                       |
cL.lntrade#c.educshare |  -.0477285   .0174113    -2.74   0.006    -.0818539    -.013603
                       |
               reg_pop |
                   L1. |   .4604118   .0722552     6.37   0.000     .3187942    .6020294
                       |
        reg_urbanshare |
                   L1. |   .0167301   .0105232     1.59   0.112    -.0038949    .0373551
                       |
               lngrppc |
                   L1. |  -.6059637   .1288986    -4.70   0.000    -.8586003   -.3533272
                       |
             reg_grpgr |
                   L1. |  -.4506051   .2050263    -2.20   0.028    -.8524493   -.0487609
                       |
     reg_levelofunempl |
                   L1. |  -.0329513   .0290634    -1.13   0.257    -.0899145    .0240118
                       |
                lnnews |
                   L1. |  -.0436215   .1127103    -0.39   0.699    -.2645295    .1772865
                       |
             lndistmos |
                   L1. |   .2066014   .0763585     2.71   0.007     .0569414    .3562614
                       |
             lnroadden |
                   L1. |  -.1243937   .0995465    -1.25   0.211    -.3195013    .0707139
                       |
           reg_density |   .0432779   .0845487     0.51   0.609    -.1224345    .2089902
                       |
                 lnfdi |
                   L1. |   .1066998   .0977078     1.09   0.275    -.0848039    .2982034
                       |
                 _cons |   4.752546   2.128839     2.23   0.026     .5800993    8.924993
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3721707   .1911184                     -.0024146    .7467559
                 /ln_s |   1.190773   .2705402                      .6605236    1.721022
-----------------------+----------------------------------------------------------------
                     r |   1.450881     .27729                      .9975883    2.110143
                     s |   3.289622   .8899748                      1.935806    5.590236
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 132.59               Prob >= chibar2 = 0.000

. est store ikd2

. * IKD protest, conditional + extended controls
. xtnbreg protest_ikd2 $X2 $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -74476.476  (not concave)
Iteration 1:   log likelihood = -62576.974  
Iteration 2:   log likelihood = -50630.648  (backed up)
Iteration 3:   log likelihood =  -36764.26  
Iteration 4:   log likelihood = -23441.835  
Iteration 5:   log likelihood = -6896.7262  
Iteration 6:   log likelihood =  -3446.534  
Iteration 7:   log likelihood = -2650.3626  
Iteration 8:   log likelihood = -2600.9612  
Iteration 9:   log likelihood =  -2600.016  
Iteration 10:  log likelihood = -2600.0137  
Iteration 11:  log likelihood = -2600.0137  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1390.6747  (not concave)
Iteration 2:   log likelihood = -1369.2858  
Iteration 3:   log likelihood = -1286.6513  
Iteration 4:   log likelihood = -1280.9839  
Iteration 5:   log likelihood =  -1280.922  
Iteration 6:   log likelihood =  -1280.922  

Fitting full model:

Iteration 0:   log likelihood = -1393.8367  (not concave)
Iteration 1:   log likelihood = -1376.1675  (not concave)
Iteration 2:   log likelihood =  -1266.408  
Iteration 3:   log likelihood = -1262.2387  
Iteration 4:   log likelihood = -1230.4718  
Iteration 5:   log likelihood = -1219.6794  
Iteration 6:   log likelihood = -1218.5542  
Iteration 7:   log likelihood = -1218.4897  
Iteration 8:   log likelihood = -1218.4895  
Iteration 9:   log likelihood = -1218.4895  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      91.42
Log likelihood  = -1218.4895                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5774318   .2440631     2.37   0.018     .0990768    1.055787
                       |
             educshare |   .0890607   .0571847     1.56   0.119    -.0230192    .2011406
                       |
cL.lntrade#c.educshare |  -.0524035   .0176025    -2.98   0.003    -.0869038   -.0179031
                       |
               reg_pop |
                   L1. |   .4833176   .0732753     6.60   0.000     .3397006    .6269346
                       |
        reg_urbanshare |
                   L1. |   .0136482   .0111266     1.23   0.220    -.0081596     .035456
                       |
               lngrppc |
                   L1. |  -.6144772   .1418066    -4.33   0.000    -.8924132   -.3365413
                       |
             reg_grpgr |
                   L1. |   -.428684   .2060136    -2.08   0.037    -.8324633   -.0249046
                       |
     reg_levelofunempl |
                   L1. |  -.0358944   .0295562    -1.21   0.225    -.0938236    .0220347
                       |
                lnnews |
                   L1. |  -.0669973   .1135012    -0.59   0.555    -.2894555     .155461
                       |
             lndistmos |
                   L1. |   .1791088   .0787578     2.27   0.023     .0247464    .3334712
                       |
             lnroadden |
                   L1. |  -.0786518   .1111613    -0.71   0.479     -.296524    .1392204
                       |
           reg_density |   .0703033   .0665109     1.06   0.291    -.0600555    .2006622
                       |
                 lnfdi |
                   L1. |     .07524   .0985714     0.76   0.445    -.1179563    .2684363
                       |
                 rents |   .0110966   .0095082     1.17   0.243    -.0075391    .0297322
             pressfill |   .0477562    .090624     0.53   0.598    -.1298636     .225376
      KPRFmandateshare |  -.0269951   .0109344    -2.47   0.014     -.048426   -.0055641
            pctRussian |   .0064228   .0061568     1.04   0.297    -.0056444      .01849
                 _cons |   4.593955   2.297061     2.00   0.046      .091799    9.096111
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3869403     .19155                      .0115092    .7623714
                 /ln_s |   1.200743   .2685418                      .6744109    1.727075
-----------------------+----------------------------------------------------------------
                     r |   1.472469   .2820514                      1.011576    2.143353
                     s |   3.322585   .8922528                      1.962876     5.62418
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 124.86               Prob >= chibar2 = 0.000

. est store ikd3

. * IKD protest, conditional + extended controls + lagged protest
. xtnbreg protest_ikd2 $X2 $C1 $C2 l.protest_ikd2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -68610.886  (not concave)
Iteration 1:   log likelihood = -57647.783  
Iteration 2:   log likelihood = -39041.413  (backed up)
Iteration 3:   log likelihood =  -22548.34  (backed up)
Iteration 4:   log likelihood = -15779.715  (backed up)
Iteration 5:   log likelihood = -8958.1442  
Iteration 6:   log likelihood = -6190.0489  
Iteration 7:   log likelihood = -2462.8819  
Iteration 8:   log likelihood =  -2188.635  
Iteration 9:   log likelihood = -2163.2292  
Iteration 10:  log likelihood = -2162.8368  
Iteration 11:  log likelihood = -2162.8363  
Iteration 12:  log likelihood = -2162.8363  

Iteration 0:   log likelihood = -3281.8302  
Iteration 1:   log likelihood = -1432.0317  
Iteration 2:   log likelihood = -1212.2738  
Iteration 3:   log likelihood =  -1211.356  
Iteration 4:   log likelihood = -1211.3552  
Iteration 5:   log likelihood = -1211.3552  

Iteration 0:   log likelihood = -1211.3552  (not concave)
Iteration 1:   log likelihood = -1176.6052  
Iteration 2:   log likelihood = -1090.1241  
Iteration 3:   log likelihood = -1074.0655  
Iteration 4:   log likelihood = -1069.9655  
Iteration 5:   log likelihood = -1069.9457  
Iteration 6:   log likelihood = -1069.9456  

Fitting full model:

Iteration 0:   log likelihood = -1159.0501  (not concave)
Iteration 1:   log likelihood = -1149.0354  
Iteration 2:   log likelihood = -1053.5644  
Iteration 3:   log likelihood = -1026.4738  
Iteration 4:   log likelihood = -1024.7803  
Iteration 5:   log likelihood = -1024.7198  
Iteration 6:   log likelihood = -1024.7196  
Iteration 7:   log likelihood = -1024.7196  

Random-effects negative binomial regression     Number of obs     =        353
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        4.8
                                                              max =          5

                                                Wald chi2(18)     =     134.08
Log likelihood  = -1024.7196                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5743686   .2659734     2.16   0.031     .0530703    1.095667
                       |
             educshare |   .0753287   .0627135     1.20   0.230    -.0475875    .1982449
                       |
cL.lntrade#c.educshare |  -.0435189   .0194684    -2.24   0.025    -.0816764   -.0053615
                       |
               reg_pop |
                   L1. |   .5146341   .0806676     6.38   0.000     .3565285    .6727398
                       |
        reg_urbanshare |
                   L1. |   .0313004   .0124781     2.51   0.012     .0068438     .055757
                       |
               lngrppc |
                   L1. |  -1.397555   .2003037    -6.98   0.000    -1.790143   -1.004967
                       |
             reg_grpgr |
                   L1. |  -.0920041   .2251747    -0.41   0.683    -.5333383    .3493301
                       |
     reg_levelofunempl |
                   L1. |  -.0146163   .0328243    -0.45   0.656    -.0789507    .0497182
                       |
                lnnews |
                   L1. |  -.0261317   .1246397    -0.21   0.834     -.270421    .2181576
                       |
             lndistmos |
                   L1. |   .1471923   .0973515     1.51   0.131    -.0436132    .3379978
                       |
             lnroadden |
                   L1. |  -.1534107   .1262501    -1.22   0.224    -.4008563     .094035
                       |
           reg_density |   .0381865   .0570536     0.67   0.503    -.0736366    .1500095
                       |
                 lnfdi |
                   L1. |   .1720628   .1069189     1.61   0.108    -.0374945      .38162
                       |
                 rents |   .0304599   .0107827     2.82   0.005     .0093262    .0515936
             pressfill |   .0940256   .1134268     0.83   0.407    -.1282869    .3163381
      KPRFmandateshare |  -.0297068   .0141311    -2.10   0.036    -.0574032   -.0020103
            pctRussian |   .0054586   .0065728     0.83   0.406    -.0074237     .018341
                       |
          protest_ikd2 |
                   L1. |   .0018726   .0017036     1.10   0.272    -.0014664    .0052115
                       |
                 _cons |   12.51955   2.707613     4.62   0.000     7.212729    17.82638
-----------------------+----------------------------------------------------------------
                 /ln_r |   .4230618   .2002433                      .0305921    .8155314
                 /ln_s |   1.189053   .2905913                      .6195044    1.758601
-----------------------+----------------------------------------------------------------
                     r |   1.526629   .3056971                      1.031065    2.260376
                     s |   3.283969    .954293                      1.858007    5.804314
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 90.45                Prob >= chibar2 = 0.000

. est store ikd4

. * IKD protest, conditional without onshore oil
. gen oilregion = 0

. replace oilregion = 1 if rname=="Adygea" | rname=="Astrakhan" | ///
>         rname=="Bashkortostan" | rname=="Chukotka" | rname=="Irkutsk" | ///
>         rname=="Kaliningrad" | rname=="Kalmykia" | rname=="Kamchatka" | ///
>         rname=="Khabarovsk" | rname=="Khanty-Mansi" | rname=="Kirov" | ///
>         rname=="Komi" | rname=="Krasnodar" | rname=="Krasnoyarsk" | ///
>         rname=="Novosibirsk" | rname=="Omsk" | rname=="Rostov" | ///
>         rname=="Sakha" | rname=="Sakhalin" | rname=="Saratov" | ///
>         rname=="Sverdlovsk" | rname=="Tatarstan" | rname=="Tyumen" | ///
>         rname=="Volgograd" | rname=="Yamalo-Nenets"
(625 real changes made)

. xtnbreg protest_ikd2 $X2 $C1 if sample1==1 & oilregion==0, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -45152.977  (not concave)
Iteration 1:   log likelihood = -16344.614  
Iteration 2:   log likelihood = -7231.1985  (backed up)
Iteration 3:   log likelihood = -3974.9778  
Iteration 4:   log likelihood = -2096.3258  
Iteration 5:   log likelihood = -1968.2135  
Iteration 6:   log likelihood = -1966.4817  
Iteration 7:   log likelihood = -1966.4806  
Iteration 8:   log likelihood = -1966.4806  

Iteration 0:   log likelihood =  -3251.021  
Iteration 1:   log likelihood = -1093.5329  
Iteration 2:   log likelihood = -1049.9767  
Iteration 3:   log likelihood = -978.84793  
Iteration 4:   log likelihood = -978.40066  
Iteration 5:   log likelihood = -978.40042  
Iteration 6:   log likelihood = -978.40042  

Iteration 0:   log likelihood = -978.40042  (not concave)
Iteration 1:   log likelihood = -938.79404  
Iteration 2:   log likelihood = -887.61769  
Iteration 3:   log likelihood = -878.93066  
Iteration 4:   log likelihood = -878.04607  
Iteration 5:   log likelihood = -878.04492  
Iteration 6:   log likelihood = -878.04492  

Fitting full model:

Iteration 0:   log likelihood = -982.78381  (not concave)
Iteration 1:   log likelihood = -869.93338  
Iteration 2:   log likelihood = -844.19789  
Iteration 3:   log likelihood = -832.21003  
Iteration 4:   log likelihood = -830.81201  
Iteration 5:   log likelihood = -830.74113  
Iteration 6:   log likelihood = -830.74106  

Random-effects negative binomial regression     Number of obs     =        299
Group variable: rcode                           Number of groups  =         53

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(13)     =      39.34
Log likelihood  = -830.74106                    Prob > chi2       =     0.0002

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5449169   .2989347     1.82   0.068    -.0409843    1.130818
                       |
             educshare |   .0897337   .0686167     1.31   0.191    -.0447525    .2242199
                       |
cL.lntrade#c.educshare |   -.053015   .0208548    -2.54   0.011    -.0938896   -.0121405
                       |
               reg_pop |
                   L1. |   .3182884   .1100539     2.89   0.004     .1025867    .5339901
                       |
        reg_urbanshare |
                   L1. |   .0196785   .0136754     1.44   0.150    -.0071247    .0464818
                       |
               lngrppc |
                   L1. |  -.7040361   .1929177    -3.65   0.000    -1.082148   -.3259243
                       |
             reg_grpgr |
                   L1. |  -.2021747   .2914746    -0.69   0.488    -.7734545    .3691051
                       |
     reg_levelofunempl |
                   L1. |   -.038303   .0384057    -1.00   0.319    -.1135768    .0369708
                       |
                lnnews |
                   L1. |  -.0921184   .1452683    -0.63   0.526    -.3768391    .1926023
                       |
             lndistmos |
                   L1. |   .0978286   .0957167     1.02   0.307    -.0897726    .2854299
                       |
             lnroadden |
                   L1. |  -.1051328   .1557176    -0.68   0.500    -.4103337     .200068
                       |
           reg_density |    .128076   .0837171     1.53   0.126    -.0360065    .2921584
                       |
                 lnfdi |
                   L1. |   .0196066   .1407154     0.14   0.889    -.2561905    .2954037
                       |
                 _cons |   6.769578     2.8814     2.35   0.019     1.122137    12.41702
-----------------------+----------------------------------------------------------------
                 /ln_r |   .1473344   .2193331                     -.2825505    .5772193
                 /ln_s |    .890619    .323599                      .2563765    1.524861
-----------------------+----------------------------------------------------------------
                     r |   1.158741   .2541503                      .7538585    1.781079
                     s |   2.436637   .7884935                      1.292239    4.594507
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 94.61                Prob >= chibar2 = 0.000

. est store ikd5

. * IKD economic protest, conditional
. xtnbreg protest_ikd_econ $X2 $C1 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood =  -29415.22  (not concave)
Iteration 1:   log likelihood = -10611.747  
Iteration 2:   log likelihood =  -5583.413  (backed up)
Iteration 3:   log likelihood = -3007.7251  
Iteration 4:   log likelihood = -2107.8533  
Iteration 5:   log likelihood = -1568.9718  
Iteration 6:   log likelihood = -1565.4472  
Iteration 7:   log likelihood = -1565.4435  
Iteration 8:   log likelihood = -1565.4435  

Iteration 0:   log likelihood = -1870.7195  
Iteration 1:   log likelihood = -1317.3035  
Iteration 2:   log likelihood = -1058.0556  
Iteration 3:   log likelihood = -1058.0528  
Iteration 4:   log likelihood = -1058.0528  

Iteration 0:   log likelihood = -1058.0528  (not concave)
Iteration 1:   log likelihood = -1029.7987  (not concave)
Iteration 2:   log likelihood = -1012.9664  
Iteration 3:   log likelihood = -976.20526  
Iteration 4:   log likelihood = -943.08144  
Iteration 5:   log likelihood = -942.51257  
Iteration 6:   log likelihood = -942.51145  
Iteration 7:   log likelihood = -942.51145  

Fitting full model:

Iteration 0:   log likelihood = -986.41081  (not concave)
Iteration 1:   log likelihood = -942.01616  
Iteration 2:   log likelihood = -923.47889  
Iteration 3:   log likelihood = -912.87929  
Iteration 4:   log likelihood = -912.17755  
Iteration 5:   log likelihood = -912.17275  
Iteration 6:   log likelihood = -912.17275  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =     104.04
Log likelihood  = -912.17275                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
      protest_ikd_econ |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .7786576   .2790699     2.79   0.005     .2316907    1.325625
                       |
             educshare |   .1088053   .0632289     1.72   0.085     -.015121    .2327316
                       |
cL.lntrade#c.educshare |  -.0560788   .0195181    -2.87   0.004    -.0943336   -.0178239
                       |
               reg_pop |
                   L1. |   .4915305   .0732441     6.71   0.000     .3479746    .6350864
                       |
        reg_urbanshare |
                   L1. |   .0067122   .0114065     0.59   0.556    -.0156441    .0290685
                       |
               lngrppc |
                   L1. |  -1.046562   .1612848    -6.49   0.000    -1.362674   -.7304492
                       |
             reg_grpgr |
                   L1. |  -.5417513   .2518858    -2.15   0.031    -1.035438   -.0480642
                       |
     reg_levelofunempl |
                   L1. |  -.0658873   .0376867    -1.75   0.080    -.1397518    .0079772
                       |
                lnnews |
                   L1. |  -.0110444   .1323756    -0.08   0.934    -.2704957    .2484069
                       |
             lndistmos |
                   L1. |   .1689045   .0858519     1.97   0.049     .0006378    .3371712
                       |
             lnroadden |
                   L1. |  -.2850386   .1081048    -2.64   0.008    -.4969201   -.0731572
                       |
           reg_density |   .1946515   .0857032     2.27   0.023     .0266763    .3626268
                       |
                 lnfdi |
                   L1. |   .1949543   .1190488     1.64   0.102     -.038377    .4282856
                       |
                 _cons |   10.14629   2.571176     3.95   0.000     5.106882    15.18571
-----------------------+----------------------------------------------------------------
                 /ln_r |   .6968121   .2287168                      .2485355    1.145089
                 /ln_s |   1.243357   .3290202                      .5984889    1.888224
-----------------------+----------------------------------------------------------------
                     r |   2.007343   .4591131                      1.282146     3.14272
                     s |   3.467232   1.140789                      1.819367    6.607624
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 60.68                Prob >= chibar2 = 0.000

. est store ikd_econ

. * MMA protest, conditional
. xtnbreg protest_mma $X2 $C1 if sample2==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -19796.076  (not concave)
Iteration 1:   log likelihood = -7150.2766  
Iteration 2:   log likelihood = -4355.4211  (backed up)
Iteration 3:   log likelihood = -2282.4443  
Iteration 4:   log likelihood = -1289.2636  
Iteration 5:   log likelihood = -899.70748  
Iteration 6:   log likelihood = -895.13684  
Iteration 7:   log likelihood = -895.12335  
Iteration 8:   log likelihood = -895.12335  

Iteration 0:   log likelihood = -1424.8803  
Iteration 1:   log likelihood = -1069.6442  
Iteration 2:   log likelihood = -984.58705  
Iteration 3:   log likelihood = -934.24036  
Iteration 4:   log likelihood = -934.17924  
Iteration 5:   log likelihood = -934.17923  

Iteration 0:   log likelihood = -934.17923  (not concave)
Iteration 1:   log likelihood =  -884.3536  
Iteration 2:   log likelihood = -797.89778  
Iteration 3:   log likelihood = -777.59755  
Iteration 4:   log likelihood = -762.37325  
Iteration 5:   log likelihood = -760.31662  
Iteration 6:   log likelihood = -760.28345  
Iteration 7:   log likelihood = -760.28344  

Fitting full model:

Iteration 0:   log likelihood = -781.92526  
Iteration 1:   log likelihood =  -769.2988  (not concave)
Iteration 2:   log likelihood = -748.23845  
Iteration 3:   log likelihood = -742.68303  
Iteration 4:   log likelihood = -742.27806  
Iteration 5:   log likelihood = -742.26801  
Iteration 6:   log likelihood = -742.26801  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =     259.20
Log likelihood  = -742.26801                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
           protest_mma |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5740065   .2695281     2.13   0.033     .0457411    1.102272
                       |
             educshare |   .1613034   .0579782     2.78   0.005     .0476683    .2749386
                       |
cL.lntrade#c.educshare |  -.0460389   .0184717    -2.49   0.013    -.0822428   -.0098349
                       |
               reg_pop |
                   L1. |   .3311763   .0591279     5.60   0.000     .2152878    .4470648
                       |
        reg_urbanshare |
                   L1. |   .0308718   .0098818     3.12   0.002      .011504    .0502397
                       |
               lngrppc |
                   L1. |   1.055324   .1526267     6.91   0.000     .7561809    1.354466
                       |
             reg_grpgr |
                   L1. |   .7556544   .2063052     3.66   0.000     .3513035    1.160005
                       |
     reg_levelofunempl |
                   L1. |    .159851   .0305236     5.24   0.000     .1000257    .2196762
                       |
                lnnews |
                   L1. |   .2576607   .1147599     2.25   0.025     .0327354    .4825859
                       |
             lndistmos |
                   L1. |   .3223942    .078939     4.08   0.000     .1676766    .4771118
                       |
             lnroadden |
                   L1. |   .4055699    .093636     4.33   0.000     .2220468     .589093
                       |
           reg_density |   .0972831   .0278888     3.49   0.000     .0426222    .1519441
                       |
                 lnfdi |
                   L1. |  -.2154428   .1046984    -2.06   0.040    -.4206479   -.0102377
                       |
                 _cons |   -24.0903   2.387804   -10.09   0.000    -28.77031   -19.41029
-----------------------+----------------------------------------------------------------
                 /ln_r |   1.935776   .2943692                      1.358823    2.512729
                 /ln_s |   1.891513   .3545865                      1.196536     2.58649
-----------------------+----------------------------------------------------------------
                     r |   6.929417   2.039807                      3.891609    12.33855
                     s |   6.629391   2.350692                      3.308636    13.28306
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 36.03                Prob >= chibar2 = 0.000

. est store mma

. * ICEWS protest, conditional
. xtnbreg protest_icews $X2 $C1 if sample2==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -11868.765  
Iteration 1:   log likelihood = -5519.4046  (backed up)
Iteration 2:   log likelihood = -3268.0437  
Iteration 3:   log likelihood = -2115.5179  
Iteration 4:   log likelihood = -1387.8597  
Iteration 5:   log likelihood = -1311.9475  
Iteration 6:   log likelihood = -1311.2388  
Iteration 7:   log likelihood = -1311.2384  
Iteration 8:   log likelihood = -1311.2384  

Iteration 0:   log likelihood = -1391.9675  
Iteration 1:   log likelihood = -1018.3994  
Iteration 2:   log likelihood = -846.48974  
Iteration 3:   log likelihood = -845.45094  
Iteration 4:   log likelihood = -845.45011  
Iteration 5:   log likelihood = -845.45011  

Iteration 0:   log likelihood = -845.45011  (not concave)
Iteration 1:   log likelihood = -822.12537  
Iteration 2:   log likelihood = -771.14282  
Iteration 3:   log likelihood = -760.67156  
Iteration 4:   log likelihood = -760.20201  
Iteration 5:   log likelihood = -760.19978  
Iteration 6:   log likelihood = -760.19978  

Fitting full model:

Iteration 0:   log likelihood = -762.08431  
Iteration 1:   log likelihood = -749.37368  (not concave)
Iteration 2:   log likelihood = -745.21667  (not concave)
Iteration 3:   log likelihood = -717.74141  
Iteration 4:   log likelihood = -714.95599  
Iteration 5:   log likelihood = -714.92415  
Iteration 6:   log likelihood = -714.92414  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =     120.34
Log likelihood  = -714.92414                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
         protest_icews |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5426997   .3516498     1.54   0.123    -.1465211    1.231921
                       |
             educshare |   .1246866   .0831873     1.50   0.134    -.0383574    .2877307
                       |
cL.lntrade#c.educshare |  -.0381812   .0258504    -1.48   0.140    -.0888471    .0124847
                       |
               reg_pop |
                   L1. |   .4688448   .0838331     5.59   0.000     .3045351    .6331546
                       |
        reg_urbanshare |
                   L1. |    .022826   .0134453     1.70   0.090    -.0035263    .0491783
                       |
               lngrppc |
                   L1. |   .2684373   .1984847     1.35   0.176    -.1205857    .6574602
                       |
             reg_grpgr |
                   L1. |   .0269149    .259382     0.10   0.917    -.4814645    .5352942
                       |
     reg_levelofunempl |
                   L1. |   .0393516   .0410822     0.96   0.338    -.0411681    .1198713
                       |
                lnnews |
                   L1. |   .1519962    .152816     0.99   0.320    -.1475177    .4515101
                       |
             lndistmos |
                   L1. |   .0281617   .1043119     0.27   0.787    -.1762858    .2326092
                       |
             lnroadden |
                   L1. |   .1249341   .1288146     0.97   0.332    -.1275378     .377406
                       |
           reg_density |  -.1028597   .0319316    -3.22   0.001    -.1654445   -.0402749
                       |
                 lnfdi |
                   L1. |  -.2336523   .1407955    -1.66   0.097    -.5096065    .0423019
                       |
                 _cons |  -9.637866   3.095079    -3.11   0.002    -15.70411   -3.571624
-----------------------+----------------------------------------------------------------
                 /ln_r |   .5762005   .2171353                      .1506231    1.001778
                 /ln_s |   .8041747   .2897516                       .236272    1.372077
-----------------------+----------------------------------------------------------------
                     r |   1.779265   .3863413                      1.162558    2.723119
                     s |   2.234851   .6475518                      1.266519    3.943535
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 90.55                Prob >= chibar2 = 0.000

. est store icews

. * KPRF protest, conditional
. xtnbreg protest_kprf2 $X2 $C1 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -1788.2237  
Iteration 1:   log likelihood = -1701.7862  
Iteration 2:   log likelihood = -1701.3335  
Iteration 3:   log likelihood = -1701.3335  

Iteration 0:   log likelihood = -1689.3343  
Iteration 1:   log likelihood = -1442.7988  
Iteration 2:   log likelihood = -1433.0717  
Iteration 3:   log likelihood = -1433.0512  
Iteration 4:   log likelihood = -1433.0512  

Iteration 0:   log likelihood = -1433.0512  (not concave)
Iteration 1:   log likelihood = -1406.4267  
Iteration 2:   log likelihood = -1356.8905  
Iteration 3:   log likelihood = -1337.0747  
Iteration 4:   log likelihood = -1336.9282  
Iteration 5:   log likelihood = -1336.9281  

Fitting full model:

Iteration 0:   log likelihood = -1339.5409  
Iteration 1:   log likelihood = -1285.4329  
Iteration 2:   log likelihood = -1282.2881  
Iteration 3:   log likelihood = -1276.5026  
Iteration 4:   log likelihood =  -1276.347  
Iteration 5:   log likelihood = -1276.3469  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =     101.26
Log likelihood  = -1276.3469                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
         protest_kprf2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |  -.0731165   .1785171    -0.41   0.682    -.4230037    .2767706
                       |
             educshare |   .0012034   .0394721     0.03   0.976    -.0761605    .0785673
                       |
cL.lntrade#c.educshare |   .0035366   .0125065     0.28   0.777    -.0209757    .0280489
                       |
               reg_pop |
                   L1. |   .1404621   .0434857     3.23   0.001     .0552318    .2256925
                       |
        reg_urbanshare |
                   L1. |   .0080626   .0065357     1.23   0.217    -.0047472    .0208723
                       |
               lngrppc |
                   L1. |   .2934769   .0832368     3.53   0.000     .1303358     .456618
                       |
             reg_grpgr |
                   L1. |   .0650195   .1307853     0.50   0.619    -.1913149    .3213539
                       |
     reg_levelofunempl |
                   L1. |   .0213062   .0182905     1.16   0.244    -.0145424    .0571548
                       |
                lnnews |
                   L1. |   .1843026   .0691238     2.67   0.008     .0488224    .3197828
                       |
             lndistmos |
                   L1. |   .1500128   .0565266     2.65   0.008     .0392227     .260803
                       |
             lnroadden |
                   L1. |   .2663405   .0642509     4.15   0.000     .1404112    .3922699
                       |
           reg_density |   .0103744   .0315848     0.33   0.743    -.0515307    .0722795
                       |
                 lnfdi |
                   L1. |   .0107655   .0645001     0.17   0.867    -.1156523    .1371833
                       |
                 _cons |  -5.878477   1.331895    -4.41   0.000    -8.488944   -3.268011
-----------------------+----------------------------------------------------------------
                 /ln_r |   2.230955   .2112937                      1.816827    2.645083
                 /ln_s |   2.347037   .2374061                       1.88173    2.812344
-----------------------+----------------------------------------------------------------
                     r |   9.308749    1.96688                      6.152304    14.08461
                     s |   10.45455   2.481973                       6.56485    16.64891
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 121.16               Prob >= chibar2 = 0.000

. est store kprf

. * KPRF economic protest, conditional
. xtnbreg protest_kprf_econ2 $X2 $C1 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -1065.5658  
Iteration 1:   log likelihood = -965.28008  
Iteration 2:   log likelihood = -962.67951  
Iteration 3:   log likelihood = -962.66959  
Iteration 4:   log likelihood = -962.66959  

Iteration 0:   log likelihood = -960.90672  
Iteration 1:   log likelihood = -951.73238  
Iteration 2:   log likelihood = -911.68514  
Iteration 3:   log likelihood =  -911.6799  
Iteration 4:   log likelihood =  -911.6799  

Iteration 0:   log likelihood =  -911.6799  (not concave)
Iteration 1:   log likelihood = -896.43656  
Iteration 2:   log likelihood = -877.29039  
Iteration 3:   log likelihood = -844.88657  
Iteration 4:   log likelihood = -844.49135  
Iteration 5:   log likelihood =  -844.4913  

Fitting full model:

Iteration 0:   log likelihood = -865.88827  
Iteration 1:   log likelihood = -839.67268  
Iteration 2:   log likelihood = -833.71209  
Iteration 3:   log likelihood = -830.21981  
Iteration 4:   log likelihood = -828.66502  
Iteration 5:   log likelihood = -828.60692  
Iteration 6:   log likelihood = -828.60686  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(13)     =     100.52
Log likelihood  = -828.60686                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
    protest_kprf_econ2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .0767182    .262644     0.29   0.770    -.4380545    .5914909
                       |
             educshare |  -.0592282   .0587649    -1.01   0.314    -.1744053    .0559489
                       |
cL.lntrade#c.educshare |   .0126446   .0184751     0.68   0.494    -.0235659     .048855
                       |
               reg_pop |
                   L1. |   .3133895   .0602073     5.21   0.000     .1953853    .4313937
                       |
        reg_urbanshare |
                   L1. |   .0160562   .0092701     1.73   0.083    -.0021128    .0342253
                       |
               lngrppc |
                   L1. |  -1.127771   .1474808    -7.65   0.000    -1.416828   -.8387143
                       |
             reg_grpgr |
                   L1. |   -.238301   .2405684    -0.99   0.322    -.7098063    .2332044
                       |
     reg_levelofunempl |
                   L1. |  -.0318235   .0314909    -1.01   0.312    -.0935445    .0298974
                       |
                lnnews |
                   L1. |   .0388901    .113479     0.34   0.732    -.1835246    .2613048
                       |
             lndistmos |
                   L1. |   .1669926   .0762043     2.19   0.028     .0176349    .3163503
                       |
             lnroadden |
                   L1. |  -.0811418   .0961644    -0.84   0.399    -.2696206     .107337
                       |
           reg_density |   .0676604   .0631578     1.07   0.284    -.0561266    .1914474
                       |
                 lnfdi |
                   L1. |   .0402794   .1061337     0.38   0.704    -.1677388    .2482976
                       |
                 _cons |   11.82945   2.255083     5.25   0.000     7.409573    16.24933
-----------------------+----------------------------------------------------------------
                 /ln_r |   2.048325   .2754103                      1.508531    2.588119
                 /ln_s |   1.862849   .3299715                      1.216116    2.509581
-----------------------+----------------------------------------------------------------
                     r |   7.754903    2.13578                      4.520086    13.30473
                     s |   6.442062   2.125697                      3.374059    12.29977
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 31.77                Prob >= chibar2 = 0.000

. est store kprf_econ

. * Table 1 in Main Text
. esttab ikd1 ikd2 ikd3 ikd4 ikd5 ikd_econ mma icews kprf, ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g p ll aic, labels("# of observations" "# of regions" ///
>                 "Prob > Chi2" "Log likelihood" "AIC") fmt(0 0 3 2 2)) ///
>         order(*lntrade* educ*)

------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
                              (1)             (2)             (3)             (4)         
>     (5)             (6)             (7)             (8)             (9)   
                     protest_ikd2    protest_ikd2    protest_ikd2    protest_ikd2    prote
> st_ikd2    (sum) prot~n     mma_protest    protest_ic~s    protest_k~f2   
                             b/se            b/se            b/se            b/se         
>    b/se            b/se            b/se            b/se            b/se   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
main                                                                                      
>                                                                           
L.Imports+Exports ~n       -0.062           0.524**         0.577**         0.574**       
>   0.545*          0.779***        0.574**         0.543          -0.073   
                           (0.11)          (0.24)          (0.24)          (0.27)         
>  (0.30)          (0.28)          (0.27)          (0.35)          (0.18)   
L.Imports+Exports ~#                       -0.048***       -0.052***       -0.044**       
>  -0.053**        -0.056***       -0.046**        -0.038           0.004   
                                           (0.02)          (0.02)          (0.02)         
>  (0.02)          (0.02)          (0.02)          (0.03)          (0.01)   
educshare                  -0.074***        0.072           0.089           0.075         
>   0.090           0.109*          0.161***        0.125           0.001   
                           (0.02)          (0.06)          (0.06)          (0.06)         
>  (0.07)          (0.06)          (0.06)          (0.08)          (0.04)   
L.reg_pop                   0.423***        0.460***        0.483***        0.515***      
>   0.318***        0.492***        0.331***        0.469***        0.140***
                           (0.07)          (0.07)          (0.07)          (0.08)         
>  (0.11)          (0.07)          (0.06)          (0.08)          (0.04)   
L.reg_urbanshare            0.008           0.017           0.014           0.031**       
>   0.020           0.007           0.031***        0.023*          0.008   
                           (0.01)          (0.01)          (0.01)          (0.01)         
>  (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   
L.lngrppc                  -0.601***       -0.606***       -0.614***       -1.398***      
>  -0.704***       -1.047***        1.055***        0.268           0.293***
                           (0.13)          (0.13)          (0.14)          (0.20)         
>  (0.19)          (0.16)          (0.15)          (0.20)          (0.08)   
L.reg_grpgr                -0.458**        -0.451**        -0.429**        -0.092         
>  -0.202          -0.542**         0.756***        0.027           0.065   
                           (0.21)          (0.21)          (0.21)          (0.23)         
>  (0.29)          (0.25)          (0.21)          (0.26)          (0.13)   
L.reg_levelofunempl        -0.031          -0.033          -0.036          -0.015         
>  -0.038          -0.066*          0.160***        0.039           0.021   
                           (0.03)          (0.03)          (0.03)          (0.03)         
>  (0.04)          (0.04)          (0.03)          (0.04)          (0.02)   
L.lnnews                   -0.103          -0.044          -0.067          -0.026         
>  -0.092          -0.011           0.258**         0.152           0.184***
                           (0.11)          (0.11)          (0.11)          (0.12)         
>  (0.15)          (0.13)          (0.11)          (0.15)          (0.07)   
L.lndistmos                 0.219***        0.207***        0.179**         0.147         
>   0.098           0.169**         0.322***        0.028           0.150***
                           (0.08)          (0.08)          (0.08)          (0.10)         
>  (0.10)          (0.09)          (0.08)          (0.10)          (0.06)   
L.lnroadden                -0.114          -0.124          -0.079          -0.153         
>  -0.105          -0.285***        0.406***        0.125           0.266***
                           (0.10)          (0.10)          (0.11)          (0.13)         
>  (0.16)          (0.11)          (0.09)          (0.13)          (0.06)   
reg_density                -0.015           0.043           0.070           0.038         
>   0.128           0.195**         0.097***       -0.103***        0.010   
                           (0.12)          (0.08)          (0.07)          (0.06)         
>  (0.08)          (0.09)          (0.03)          (0.03)          (0.03)   
L.FDI inflows (% o~n        0.114           0.107           0.075           0.172         
>   0.020           0.195          -0.215**        -0.234*          0.011   
                           (0.10)          (0.10)          (0.10)          (0.11)         
>  (0.14)          (0.12)          (0.10)          (0.14)          (0.06)   
(mean) rents                                                0.011           0.030***      
>                                                                           
                                                           (0.01)          (0.01)         
>                                                                           
(mean) pressfill                                            0.048           0.094         
>                                                                           
                                                           (0.09)          (0.11)         
>                                                                           
(mean) KPRFmandate~e                                       -0.027**        -0.030**       
>                                                                           
                                                           (0.01)          (0.01)         
>                                                                           
(mean) pctRussian                                           0.006           0.005         
>                                                                           
                                                           (0.01)          (0.01)         
>                                                                           
L.protest_ikd2                                                              0.002         
>                                                                           
                                                                           (0.00)         
>                                                                           
Constant                    7.475***        4.753**         4.594**        12.520***      
>   6.770**        10.146***      -24.090***       -9.638***       -5.878***
                           (1.94)          (2.13)          (2.30)          (2.71)         
>  (2.88)          (2.57)          (2.39)          (3.10)          (1.33)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
ln_r                                                                                      
>                                                                           
Constant                    0.291           0.372*          0.387**         0.423**       
>   0.147           0.697***        1.936***        0.576***        2.231***
                           (0.19)          (0.19)          (0.19)          (0.20)         
>  (0.22)          (0.23)          (0.29)          (0.22)          (0.21)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
ln_s                                                                                      
>                                                                           
Constant                    1.059***        1.191***        1.201***        1.189***      
>   0.891***        1.243***        1.892***        0.804***        2.347***
                           (0.26)          (0.27)          (0.27)          (0.29)         
>  (0.32)          (0.33)          (0.35)          (0.29)          (0.24)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
# of observations             427             427             416             353         
>     299             427             427             427             427   
# of regions                   75              75              74              74         
>      53              75              75              75              75   
Prob > Chi2                 0.000           0.000           0.000           0.000         
>   0.000           0.000           0.000           0.000           0.000   
Log likelihood           -1252.97        -1249.40        -1218.49        -1024.72         
> -830.74         -912.17         -742.27         -714.92        -1276.35   
AIC                       2535.95         2530.80         2476.98         2091.44         
> 1693.48         1856.35         1516.54         1461.85         2584.69   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------

. 
. *** Additional control variables
. * IKD protest, including FDI interaction
. xtnbreg protest_ikd2 $X2 $C1 cl.lnfdi##c.educshare  if sample1==1, re
note: L.lnfdi omitted because of collinearity
note: educshare omitted because of collinearity

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76887.229  (not concave)
Iteration 1:   log likelihood = -27738.927  
Iteration 2:   log likelihood = -12250.026  (backed up)
Iteration 3:   log likelihood = -9162.4611  
Iteration 4:   log likelihood =  -2965.624  
Iteration 5:   log likelihood = -2745.2703  
Iteration 6:   log likelihood = -2740.6214  
Iteration 7:   log likelihood = -2740.6178  
Iteration 8:   log likelihood = -2740.6178  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4231  (not concave)
Iteration 2:   log likelihood = -1397.5528  
Iteration 3:   log likelihood = -1317.1582  
Iteration 4:   log likelihood = -1312.4176  
Iteration 5:   log likelihood = -1312.3637  
Iteration 6:   log likelihood = -1312.3637  

Fitting full model:

Iteration 0:   log likelihood = -1413.5335  (not concave)
Iteration 1:   log likelihood = -1280.4791  
Iteration 2:   log likelihood = -1263.5613  (not concave)
Iteration 3:   log likelihood = -1246.2205  
Iteration 4:   log likelihood = -1245.4277  
Iteration 5:   log likelihood =  -1245.417  
Iteration 6:   log likelihood =  -1245.417  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      83.86
Log likelihood  =  -1245.417                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .4154821   .2437357     1.70   0.088    -.0622311    .8931953
                       |
             educshare |   .0817989   .0558033     1.47   0.143    -.0275735    .1911712
                       |
cL.lntrade#c.educshare |  -.0408956   .0176274    -2.32   0.020    -.0754446   -.0063465
                       |
               reg_pop |
                   L1. |   .4743634   .0706903     6.71   0.000     .3358129    .6129139
                       |
        reg_urbanshare |
                   L1. |   .0172257   .0105411     1.63   0.102    -.0034344    .0378859
                       |
               lngrppc |
                   L1. |    -.64443   .1288074    -5.00   0.000    -.8968879   -.3919721
                       |
             reg_grpgr |
                   L1. |  -.4160273   .2071542    -2.01   0.045    -.8220421   -.0100126
                       |
     reg_levelofunempl |
                   L1. |  -.0279156   .0289672    -0.96   0.335    -.0846902     .028859
                       |
                lnnews |
                   L1. |  -.0476885   .1116942    -0.43   0.669    -.2666053    .1712282
                       |
             lndistmos |
                   L1. |    .184556   .0768259     2.40   0.016     .0339799     .335132
                       |
             lnroadden |
                   L1. |  -.1455613   .1009678    -1.44   0.149    -.3434546    .0523321
                       |
           reg_density |   .0609577   .0822117     0.74   0.458    -.1001743    .2220897
                       |
                 lnfdi |
                   L1. |   .7088584   .2330663     3.04   0.002     .2520568     1.16566
                   L1. |          0  (omitted)
                       |
             educshare |          0  (omitted)
                       |
  cL.lnfdi#c.educshare |  -.0506202   .0185135    -2.73   0.006     -.086906   -.0143344
                       |
                 _cons |   5.360716   2.147358     2.50   0.013     1.151971    9.569461
-----------------------+----------------------------------------------------------------
                 /ln_r |     .36638   .1903094                     -.0066196    .7393796
                 /ln_s |   1.166292   .2696849                      .6377193    1.694865
-----------------------+----------------------------------------------------------------
                     r |   1.442503   .2745219                      .9934023    2.094636
                     s |   3.210067   .8657066                       1.89216    5.445908
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 133.89               Prob >= chibar2 = 0.000

. est store control1

. * IKD protest, including cost of living
. xtnbreg protest_ikd2 $X2 $C1 reg_costliving if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -77084.323  (not concave)
Iteration 1:   log likelihood = -27809.328  
Iteration 2:   log likelihood = -20888.175  (backed up)
Iteration 3:   log likelihood = -12895.785  
Iteration 4:   log likelihood = -6273.0592  
Iteration 5:   log likelihood = -2874.1652  
Iteration 6:   log likelihood =  -2732.012  
Iteration 7:   log likelihood = -2728.6006  
Iteration 8:   log likelihood = -2728.5944  
Iteration 9:   log likelihood = -2728.5944  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4939  (not concave)
Iteration 2:   log likelihood = -1397.6334  
Iteration 3:   log likelihood = -1318.0795  
Iteration 4:   log likelihood = -1314.0256  
Iteration 5:   log likelihood = -1313.9892  
Iteration 6:   log likelihood = -1313.9892  

Fitting full model:

Iteration 0:   log likelihood = -1416.0366  (not concave)
Iteration 1:   log likelihood = -1404.4502  
Iteration 2:   log likelihood = -1356.1582  
Iteration 3:   log likelihood = -1288.5904  
Iteration 4:   log likelihood = -1279.0518  
Iteration 5:   log likelihood = -1249.4624  
Iteration 6:   log likelihood = -1244.7467  
Iteration 7:   log likelihood = -1244.6618  
Iteration 8:   log likelihood = -1244.6617  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      83.83
Log likelihood  = -1244.6617                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |    .379278   .2504262     1.51   0.130    -.1115482    .8701043
                       |
             educshare |   .0701592   .0565724     1.24   0.215    -.0407207    .1810392
                       |
cL.lntrade#c.educshare |  -.0424193   .0179424    -2.36   0.018    -.0775858   -.0072528
                       |
               reg_pop |
                   L1. |   .4094569   .0749738     5.46   0.000     .2625108    .5564029
                       |
        reg_urbanshare |
                   L1. |   .0215641    .010844     1.99   0.047     .0003102    .0428179
                       |
               lngrppc |
                   L1. |   .0062412    .235257     0.03   0.979    -.4548541    .4673366
                       |
             reg_grpgr |
                   L1. |  -.6589329   .2124807    -3.10   0.002    -1.075387   -.2424784
                       |
     reg_levelofunempl |
                   L1. |  -.0126645   .0297814    -0.43   0.671    -.0710349    .0457059
                       |
                lnnews |
                   L1. |  -.0122105    .114537    -0.11   0.915     -.236699     .212278
                       |
             lndistmos |
                   L1. |    .207584   .0770363     2.69   0.007     .0565956    .3585724
                       |
             lnroadden |
                   L1. |  -.1248631   .0995957    -1.25   0.210     -.320067    .0703408
                       |
           reg_density |  -.0704864   .1181562    -0.60   0.551    -.3020683    .1610955
                       |
                 lnfdi |
                   L1. |   .1083653   .0990994     1.09   0.274    -.0858659    .3025965
                       |
        reg_costliving |  -.0002294   .0000743    -3.09   0.002    -.0003749   -.0000839
                 _cons |  -1.596874   2.958853    -0.54   0.589    -7.396119    4.202371
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3692512   .1862587                      .0041908    .7343116
                 /ln_s |   1.170996   .2597755                      .6618454    1.680147
-----------------------+----------------------------------------------------------------
                     r |   1.446651   .2694514                        1.0042    2.084047
                     s |   3.225204    .837829                      1.938366    5.366344
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 138.65               Prob >= chibar2 = 0.000

. est store control2

. * IKD protest, including % of people living below subsistence
. xtnbreg protest_ikd2 $X2 $C1 reg_belowcost if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76981.267  (not concave)
Iteration 1:   log likelihood = -27766.923  
Iteration 2:   log likelihood = -11349.537  (backed up)
Iteration 3:   log likelihood =  -5371.377  
Iteration 4:   log likelihood = -2852.9747  
Iteration 5:   log likelihood = -2691.4505  
Iteration 6:   log likelihood =  -2690.373  
Iteration 7:   log likelihood = -2690.3724  
Iteration 8:   log likelihood = -2690.3724  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood =  -1419.504  (not concave)
Iteration 2:   log likelihood = -1397.7569  
Iteration 3:   log likelihood = -1317.4074  
Iteration 4:   log likelihood =   -1310.73  
Iteration 5:   log likelihood = -1310.6194  
Iteration 6:   log likelihood = -1310.6193  

Fitting full model:

Iteration 0:   log likelihood = -1409.4814  (not concave)
Iteration 1:   log likelihood = -1322.0243  
Iteration 2:   log likelihood = -1279.1033  
Iteration 3:   log likelihood = -1250.8048  
Iteration 4:   log likelihood =  -1247.228  
Iteration 5:   log likelihood = -1247.1351  
Iteration 6:   log likelihood = -1247.1349  
Iteration 7:   log likelihood = -1247.1349  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      79.40
Log likelihood  = -1247.1349                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5854047   .2415575     2.42   0.015     .1119606    1.058849
                       |
             educshare |   .0814543   .0559664     1.46   0.146    -.0282378    .1911465
                       |
cL.lntrade#c.educshare |  -.0512349   .0174312    -2.94   0.003    -.0853994   -.0170704
                       |
               reg_pop |
                   L1. |   .4773163   .0720483     6.62   0.000     .3361043    .6185283
                       |
        reg_urbanshare |
                   L1. |   .0188994   .0105688     1.79   0.074    -.0018152    .0396139
                       |
               lngrppc |
                   L1. |  -.4096631   .1579527    -2.59   0.009    -.7192447   -.1000814
                       |
             reg_grpgr |
                   L1. |  -.4894715   .2024544    -2.42   0.016    -.8862747   -.0926682
                       |
     reg_levelofunempl |
                   L1. |   -.041855    .029054    -1.44   0.150    -.0987998    .0150898
                       |
                lnnews |
                   L1. |  -.0666228   .1132782    -0.59   0.556     -.288644    .1553984
                       |
             lndistmos |
                   L1. |   .2143438   .0763587     2.81   0.005     .0646836    .3640041
                       |
             lnroadden |
                   L1. |  -.0504454   .1057041    -0.48   0.633    -.2576217    .1567308
                       |
           reg_density |   .0546517   .0724843     0.75   0.451     -.087415    .1967184
                       |
                 lnfdi |
                   L1. |     .07486   .0992561     0.75   0.451    -.1196784    .2693985
                       |
         reg_belowcost |   .0428288   .0198348     2.16   0.031     .0039533    .0817044
                 _cons |   1.247517   2.691295     0.46   0.643    -4.027325    6.522359
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3822377   .1939296                      .0021428    .7623326
                 /ln_s |   1.190952   .2754162                       .651146    1.730758
-----------------------+----------------------------------------------------------------
                     r |    1.46556   .2842155                      1.002145     2.14327
                     s |   3.290212   .9061777                      1.917737     5.64493
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 126.97               Prob >= chibar2 = 0.000

. est store control3

. * IKD protest, including cost of a fixed basket of goods
. xtnbreg protest_ikd2 $X2 $C1 reg_gdsfixed if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76943.926  (not concave)
Iteration 1:   log likelihood = -27759.377  
Iteration 2:   log likelihood = -27642.331  (backed up)
Iteration 3:   log likelihood = -19818.234  
Iteration 4:   log likelihood = -8923.2956  
Iteration 5:   log likelihood = -3134.4734  
Iteration 6:   log likelihood = -2760.2187  
Iteration 7:   log likelihood = -2731.6153  
Iteration 8:   log likelihood = -2731.5316  
Iteration 9:   log likelihood = -2731.5316  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4949  (not concave)
Iteration 2:   log likelihood = -1397.6261  
Iteration 3:   log likelihood =  -1318.024  
Iteration 4:   log likelihood = -1313.8102  
Iteration 5:   log likelihood = -1313.7704  
Iteration 6:   log likelihood = -1313.7704  

Fitting full model:

Iteration 0:   log likelihood = -1413.1146  (not concave)
Iteration 1:   log likelihood = -1386.4645  
Iteration 2:   log likelihood = -1296.7245  
Iteration 3:   log likelihood = -1262.8636  
Iteration 4:   log likelihood = -1251.3844  
Iteration 5:   log likelihood = -1242.2089  
Iteration 6:   log likelihood =  -1241.784  
Iteration 7:   log likelihood = -1241.7826  
Iteration 8:   log likelihood = -1241.7826  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      90.42
Log likelihood  = -1241.7826                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .3764852   .2469327     1.52   0.127     -.107494    .8604643
                       |
             educshare |   .0790242   .0563811     1.40   0.161    -.0314808    .1895291
                       |
cL.lntrade#c.educshare |  -.0429015   .0178256    -2.41   0.016    -.0778391   -.0079639
                       |
               reg_pop |
                   L1. |   .3858629   .0765229     5.04   0.000     .2358808    .5358451
                       |
        reg_urbanshare |
                   L1. |   .0212869   .0108548     1.96   0.050     .0000119    .0425619
                       |
               lngrppc |
                   L1. |   .3346658   .2661862     1.26   0.209    -.1870497    .8563812
                       |
             reg_grpgr |
                   L1. |  -.7554357    .214964    -3.51   0.000    -1.176757    -.334114
                       |
     reg_levelofunempl |
                   L1. |  -.0048789   .0296227    -0.16   0.869    -.0629384    .0531806
                       |
                lnnews |
                   L1. |  -.0143876   .1148657    -0.13   0.900    -.2395202     .210745
                       |
             lndistmos |
                   L1. |   .2148393   .0765157     2.81   0.005     .0648713    .3648074
                       |
             lnroadden |
                   L1. |  -.0611488   .1016286    -0.60   0.547    -.2603371    .1380396
                       |
           reg_density |   -.078424   .1100971    -0.71   0.476    -.2942104    .1373623
                       |
                 lnfdi |
                   L1. |   .0864579   .1000096     0.86   0.387    -.1095573    .2824732
                       |
          reg_gdsfixed |  -.0002383   .0000596    -4.00   0.000     -.000355   -.0001215
                 _cons |  -5.228836   3.265233    -1.60   0.109    -11.62857    1.170903
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3671449   .1844189                      .0056905    .7285993
                 /ln_s |   1.143808   .2549537                      .6441075    1.643508
-----------------------+----------------------------------------------------------------
                     r |   1.443607   .2662284                      1.005707    2.072176
                     s |   3.138696   .8002223                      1.904287    5.173284
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 143.98               Prob >= chibar2 = 0.000

. est store control4

. * IKD protest, including economic inequality
. xtnbreg protest_ikd2 $X2 $C1 reg_mincgini if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood =  -76794.81  (not concave)
Iteration 1:   log likelihood = -27705.552  
Iteration 2:   log likelihood = -17135.717  (backed up)
Iteration 3:   log likelihood = -9267.2833  
Iteration 4:   log likelihood = -5144.1155  
Iteration 5:   log likelihood = -2802.9226  
Iteration 6:   log likelihood = -2742.4977  
Iteration 7:   log likelihood = -2741.9526  
Iteration 8:   log likelihood = -2741.9525  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.4797  (not concave)
Iteration 2:   log likelihood = -1397.7148  
Iteration 3:   log likelihood = -1318.4837  
Iteration 4:   log likelihood = -1314.8617  
Iteration 5:   log likelihood = -1314.8313  
Iteration 6:   log likelihood = -1314.8313  

Fitting full model:

Iteration 0:   log likelihood =  -1418.742  (not concave)
Iteration 1:   log likelihood = -1400.6618  (not concave)
Iteration 2:   log likelihood = -1298.2657  (not concave)
Iteration 3:   log likelihood = -1260.4438  
Iteration 4:   log likelihood = -1249.6383  
Iteration 5:   log likelihood = -1248.6509  
Iteration 6:   log likelihood = -1248.5927  
Iteration 7:   log likelihood = -1248.5924  
Iteration 8:   log likelihood = -1248.5924  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      76.29
Log likelihood  = -1248.5924                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5324085   .2409743     2.21   0.027     .0601076    1.004709
                       |
             educshare |   .0740129   .0558417     1.33   0.185    -.0354349    .1834608
                       |
cL.lntrade#c.educshare |  -.0485812   .0174733    -2.78   0.005    -.0828282   -.0143343
                       |
               reg_pop |
                   L1. |   .4069151   .0846658     4.81   0.000     .2409732     .572857
                       |
        reg_urbanshare |
                   L1. |   .0180134   .0105983     1.70   0.089    -.0027589    .0387856
                       |
               lngrppc |
                   L1. |  -.6627325   .1376546    -4.81   0.000    -.9325305   -.3929345
                       |
             reg_grpgr |
                   L1. |  -.4664789   .2036879    -2.29   0.022    -.8656998    -.067258
                       |
     reg_levelofunempl |
                   L1. |  -.0317663   .0290799    -1.09   0.275    -.0887619    .0252293
                       |
                lnnews |
                   L1. |   -.034969   .1127486    -0.31   0.756    -.2559523    .1860142
                       |
             lndistmos |
                   L1. |   .1899162    .076623     2.48   0.013     .0397379    .3400945
                       |
             lnroadden |
                   L1. |  -.1084252   .0997542    -1.09   0.277    -.3039398    .0870893
                       |
           reg_density |  -.0073859   .1011373    -0.07   0.942    -.2056113    .1908395
                       |
                 lnfdi |
                   L1. |   .1178021   .0981917     1.20   0.230    -.0746501    .3102542
                       |
          reg_mincgini |   5.386883   4.300779     1.25   0.210    -3.042489    13.81626
                 _cons |    3.29618   2.419989     1.36   0.173    -1.446911    8.039272
-----------------------+----------------------------------------------------------------
                 /ln_r |    .375226   .1904585                      .0019342    .7485177
                 /ln_s |    1.19289   .2696098                      .6644649    1.721316
-----------------------+----------------------------------------------------------------
                     r |    1.45532   .2771781                      1.001936    2.113864
                     s |   3.296596   .8887947                       1.94345    5.591883
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 132.48               Prob >= chibar2 = 0.000

. est store control5

. * IKD protest, including interregional migration
. xtnbreg protest_ikd2 $X2 $C1 reg_interregmigration if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76790.592  (not concave)
Iteration 1:   log likelihood = -27704.546  
Iteration 2:   log likelihood = -12249.722  (backed up)
Iteration 3:   log likelihood =  -8347.175  
Iteration 4:   log likelihood = -2957.0325  
Iteration 5:   log likelihood = -2739.8426  
Iteration 6:   log likelihood = -2736.3572  
Iteration 7:   log likelihood = -2736.3548  
Iteration 8:   log likelihood = -2736.3548  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1421.2298  (not concave)
Iteration 2:   log likelihood = -1407.4714  
Iteration 3:   log likelihood = -1320.6947  
Iteration 4:   log likelihood = -1315.3325  
Iteration 5:   log likelihood =  -1315.281  
Iteration 6:   log likelihood =  -1315.281  

Fitting full model:

Iteration 0:   log likelihood = -1413.2365  (not concave)
Iteration 1:   log likelihood = -1364.2037  
Iteration 2:   log likelihood = -1275.0424  
Iteration 3:   log likelihood = -1250.1901  
Iteration 4:   log likelihood = -1247.2579  
Iteration 5:   log likelihood = -1247.2106  
Iteration 6:   log likelihood = -1247.2106  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      79.23
Log likelihood  = -1247.2106                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .4588967    .240363     1.91   0.056    -.0122061    .9299995
                       |
             educshare |   .0585024   .0555958     1.05   0.293    -.0504634    .1674683
                       |
cL.lntrade#c.educshare |  -.0397089    .017766    -2.24   0.025    -.0745297   -.0048882
                       |
               reg_pop |
                   L1. |   .5122287   .0752755     6.80   0.000     .3646915    .6597659
                       |
        reg_urbanshare |
                   L1. |   .0182766   .0104799     1.74   0.081    -.0022636    .0388168
                       |
               lngrppc |
                   L1. |  -.5788938   .1281325    -4.52   0.000     -.830029   -.3277587
                       |
             reg_grpgr |
                   L1. |   -.500963   .2042211    -2.45   0.014     -.901229   -.1006971
                       |
     reg_levelofunempl |
                   L1. |  -.0386642   .0291517    -1.33   0.185    -.0958004     .018472
                       |
                lnnews |
                   L1. |  -.0159834   .1120182    -0.14   0.887     -.235535    .2035682
                       |
             lndistmos |
                   L1. |   .1713093   .0754052     2.27   0.023     .0235179    .3191007
                       |
             lnroadden |
                   L1. |  -.1365462   .0978537    -1.40   0.163     -.328336    .0552435
                       |
           reg_density |  -.0093808    .087479    -0.11   0.915    -.1808365    .1620749
                       |
                 lnfdi |
                   L1. |   .1231329    .096736     1.27   0.203    -.0664662     .312732
                       |
 reg_interregmigration |  -.0000154   7.95e-06    -1.94   0.052     -.000031    1.40e-07
                 _cons |   4.442839   2.121315     2.09   0.036     .2851387     8.60054
-----------------------+----------------------------------------------------------------
                 /ln_r |   .4067167   .1896138                      .0350804    .7783529
                 /ln_s |    1.24504   .2660042                      .7236809    1.766398
-----------------------+----------------------------------------------------------------
                     r |   1.501878   .2847769                      1.035703    2.177882
                     s |   3.473072   .9238518                      2.062009    5.849746
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 136.14               Prob >= chibar2 = 0.000

. est store control6

. * IKD protest, including infant mortality
. xtnbreg protest_ikd2 $X2 $C1 reg_infantmort if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76840.982  (not concave)
Iteration 1:   log likelihood = -27722.135  
Iteration 2:   log likelihood = -10851.386  (backed up)
Iteration 3:   log likelihood = -4902.0174  
Iteration 4:   log likelihood = -2882.4461  
Iteration 5:   log likelihood = -2692.4786  
Iteration 6:   log likelihood = -2691.0948  
Iteration 7:   log likelihood = -2691.0938  
Iteration 8:   log likelihood = -2691.0938  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.5331  (not concave)
Iteration 2:   log likelihood = -1397.7088  
Iteration 3:   log likelihood = -1314.5238  
Iteration 4:   log likelihood = -1308.5647  
Iteration 5:   log likelihood = -1308.4687  
Iteration 6:   log likelihood = -1308.4687  

Fitting full model:

Iteration 0:   log likelihood =   -1419.42  (not concave)
Iteration 1:   log likelihood = -1344.3566  
Iteration 2:   log likelihood = -1258.6851  
Iteration 3:   log likelihood = -1241.3643  
Iteration 4:   log likelihood = -1240.6528  
Iteration 5:   log likelihood = -1240.6473  
Iteration 6:   log likelihood = -1240.6473  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      90.31
Log likelihood  = -1240.6473                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |    .603809   .2382332     2.53   0.011     .1368805    1.070737
                       |
             educshare |     .10429   .0561604     1.86   0.063    -.0057824    .2143624
                       |
cL.lntrade#c.educshare |  -.0579772   .0176781    -3.28   0.001    -.0926255   -.0233288
                       |
               reg_pop |
                   L1. |   .5461534   .0743647     7.34   0.000     .4004013    .6919055
                       |
        reg_urbanshare |
                   L1. |   .0125387     .01043     1.20   0.229    -.0079036    .0329811
                       |
               lngrppc |
                   L1. |  -.8386998    .142398    -5.89   0.000    -1.117795   -.5596048
                       |
             reg_grpgr |
                   L1. |  -.2691033   .2036436    -1.32   0.186    -.6682374    .1300307
                       |
     reg_levelofunempl |
                   L1. |  -.0417693   .0280501    -1.49   0.136    -.0967464    .0132078
                       |
                lnnews |
                   L1. |  -.0687491   .1120351    -0.61   0.539    -.2883338    .1508356
                       |
             lndistmos |
                   L1. |   .2361635   .0772568     3.06   0.002     .0847431     .387584
                       |
             lnroadden |
                   L1. |  -.2495982   .1049156    -2.38   0.017     -.455229   -.0439675
                       |
           reg_density |  -.0237547   .1081953    -0.22   0.826    -.2358136    .1883041
                       |
                 lnfdi |
                   L1. |   .1160106   .0986111     1.18   0.239    -.0772636    .3092848
                       |
        reg_infantmort |   -.140441   .0340596    -4.12   0.000    -.2071965   -.0736855
                 _cons |   9.096782   2.390541     3.81   0.000     4.411408    13.78216
-----------------------+----------------------------------------------------------------
                 /ln_r |     .41012   .1894302                      .0388437    .7813964
                 /ln_s |   1.259058   .2655499                      .7385895    1.779526
-----------------------+----------------------------------------------------------------
                     r |   1.506999    .285471                      1.039608    2.184521
                     s |   3.522101   .9352934                      2.092981    5.927045
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 135.64               Prob >= chibar2 = 0.000

. est store control7

. * IKD protest, including life expectancy
. xtnbreg protest_ikd2 $X2 $C1 reg_lifeexp if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76449.546  (not concave)
Iteration 1:   log likelihood = -27579.762  
Iteration 2:   log likelihood =  -13058.98  (backed up)
Iteration 3:   log likelihood = -6661.7055  
Iteration 4:   log likelihood = -2856.1108  
Iteration 5:   log likelihood = -2726.5074  
Iteration 6:   log likelihood = -2721.8862  
Iteration 7:   log likelihood =  -2721.882  
Iteration 8:   log likelihood =  -2721.882  

Iteration 0:   log likelihood = -3849.5211  
Iteration 1:   log likelihood =  -1724.977  
Iteration 2:   log likelihood = -1453.9007  
Iteration 3:   log likelihood = -1452.7576  
Iteration 4:   log likelihood = -1452.7566  
Iteration 5:   log likelihood = -1452.7566  

Iteration 0:   log likelihood = -1452.7566  (not concave)
Iteration 1:   log likelihood = -1411.4755  (not concave)
Iteration 2:   log likelihood = -1389.8648  
Iteration 3:   log likelihood = -1311.1192  
Iteration 4:   log likelihood =  -1306.183  
Iteration 5:   log likelihood = -1306.1268  
Iteration 6:   log likelihood = -1306.1268  

Fitting full model:

Iteration 0:   log likelihood = -1411.1885  (not concave)
Iteration 1:   log likelihood = -1345.3492  
Iteration 2:   log likelihood = -1257.8045  
Iteration 3:   log likelihood = -1251.1866  
Iteration 4:   log likelihood = -1241.3958  
Iteration 5:   log likelihood = -1241.0662  
Iteration 6:   log likelihood = -1241.0594  
Iteration 7:   log likelihood = -1241.0594  

Random-effects negative binomial regression     Number of obs     =        425
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      79.75
Log likelihood  = -1241.0594                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5635097   .2373394     2.37   0.018      .098333    1.028686
                       |
             educshare |    .063392   .0550341     1.15   0.249    -.0444729     .171257
                       |
cL.lntrade#c.educshare |  -.0480173   .0170971    -2.81   0.005     -.081527   -.0145076
                       |
               reg_pop |
                   L1. |   .4521223   .0706226     6.40   0.000     .3137045      .59054
                       |
        reg_urbanshare |
                   L1. |   .0172021   .0103503     1.66   0.097    -.0030842    .0374884
                       |
               lngrppc |
                   L1. |  -.8311493   .1792678    -4.64   0.000    -1.182508   -.4797909
                       |
             reg_grpgr |
                   L1. |  -.4563027   .2055265    -2.22   0.026    -.8591273   -.0534782
                       |
     reg_levelofunempl |
                   L1. |  -.0399507   .0294798    -1.36   0.175    -.0977299    .0178286
                       |
                lnnews |
                   L1. |  -.0785296   .1124825    -0.70   0.485    -.2989913    .1419321
                       |
             lndistmos |
                   L1. |   .1787056   .0771956     2.31   0.021     .0274049    .3300063
                       |
             lnroadden |
                   L1. |   -.251908   .1195423    -2.11   0.035    -.4862065   -.0176094
                       |
           reg_density |   .0396783   .0873904     0.45   0.650    -.1316038    .2109604
                       |
                 lnfdi |
                   L1. |   .1401923   .0972618     1.44   0.149    -.0504373    .3308218
                       |
           reg_lifeexp |   .0864316   .0467172     1.85   0.064    -.0051325    .1779957
                 _cons |   2.554076   2.400678     1.06   0.287    -2.151167    7.259319
-----------------------+----------------------------------------------------------------
                 /ln_r |   .4235411   .1919915                      .0472448    .7998375
                 /ln_s |     1.2948   .2714243                      .7628184    1.826782
-----------------------+----------------------------------------------------------------
                     r |   1.527361   .2932402                      1.048379    2.225179
                     s |   3.650267   .9907709                      2.144311    6.213858
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 130.13               Prob >= chibar2 = 0.000

. est store control8

. * IKD protest, including public expenditure differential index
. xtnbreg protest_ikd2 $X2 $C1 reg_ibr if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -76819.511  (not concave)
Iteration 1:   log likelihood = -27714.622  
Iteration 2:   log likelihood = -11507.702  (backed up)
Iteration 3:   log likelihood = -6115.0806  
Iteration 4:   log likelihood = -2982.9471  
Iteration 5:   log likelihood =  -2665.832  
Iteration 6:   log likelihood = -2662.4195  
Iteration 7:   log likelihood = -2662.4045  
Iteration 8:   log likelihood = -2662.4045  

Iteration 0:   log likelihood = -3854.1316  
Iteration 1:   log likelihood = -1733.5946  
Iteration 2:   log likelihood = -1462.3202  
Iteration 3:   log likelihood = -1461.0605  
Iteration 4:   log likelihood = -1461.0593  
Iteration 5:   log likelihood = -1461.0593  

Iteration 0:   log likelihood = -1461.0593  (not concave)
Iteration 1:   log likelihood = -1419.5797  (not concave)
Iteration 2:   log likelihood = -1397.6895  
Iteration 3:   log likelihood = -1316.2717  
Iteration 4:   log likelihood = -1309.6754  
Iteration 5:   log likelihood = -1309.4791  
Iteration 6:   log likelihood = -1309.4786  
Iteration 7:   log likelihood = -1309.4786  

Fitting full model:

Iteration 0:   log likelihood = -1418.4916  (not concave)
Iteration 1:   log likelihood = -1298.7378  
Iteration 2:   log likelihood = -1252.6948  
Iteration 3:   log likelihood =   -1246.74  
Iteration 4:   log likelihood = -1246.6867  
Iteration 5:   log likelihood = -1246.6866  

Random-effects negative binomial regression     Number of obs     =        427
Group variable: rcode                           Number of groups  =         75

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.7
                                                              max =          6

                                                Wald chi2(14)     =      79.68
Log likelihood  = -1246.6866                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5514185   .2393273     2.30   0.021     .0823456    1.020491
                       |
             educshare |   .0862134   .0555046     1.55   0.120    -.0225737    .1950005
                       |
cL.lntrade#c.educshare |  -.0481988   .0172028    -2.80   0.005    -.0819157    -.014482
                       |
               reg_pop |
                   L1. |   .4294984   .0726983     5.91   0.000     .2870124    .5719844
                       |
        reg_urbanshare |
                   L1. |   .0135772    .010455     1.30   0.194    -.0069142    .0340686
                       |
               lngrppc |
                   L1. |  -.5407167   .1312125    -4.12   0.000    -.7978886   -.2835449
                       |
             reg_grpgr |
                   L1. |  -.4704298   .2057625    -2.29   0.022    -.8737168   -.0671428
                       |
     reg_levelofunempl |
                   L1. |  -.0405386   .0291986    -1.39   0.165    -.0977668    .0166896
                       |
                lnnews |
                   L1. |  -.0346499   .1115956    -0.31   0.756    -.2533733    .1840735
                       |
             lndistmos |
                   L1. |    .155425   .0792778     1.96   0.050     .0000434    .3108066
                       |
             lnroadden |
                   L1. |  -.3334592   .1338455    -2.49   0.013    -.5957915   -.0711269
                       |
           reg_density |   .0290829    .092734     0.31   0.754    -.1526724    .2108382
                       |
                 lnfdi |
                   L1. |   .0924876   .0992546     0.93   0.351    -.1020477     .287023
                       |
               reg_ibr |  -.3970986   .1790128    -2.22   0.027    -.7479574   -.0462399
                 _cons |   5.731176   2.157688     2.66   0.008     1.502185    9.960166
-----------------------+----------------------------------------------------------------
                 /ln_r |   .4306912   .1947108                      .0490651    .8123173
                 /ln_s |    1.30242   .2753777                      .7626894     1.84215
-----------------------+----------------------------------------------------------------
                     r |    1.53832   .2995276                      1.050289    2.253123
                     s |   3.678186    1.01289                      2.144035     6.31009
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 125.58               Prob >= chibar2 = 0.000

. est store control9

. * Table A4 in Online Appendix
. esttab control1 control2 control3 control4 control5 control6 control7 control8 control9,
>  ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g p ll aic, labels("# of observations" "# of regions" ///
>                 "Prob > Chi2" "Log likelihood" "AIC") fmt(0 0 3 2 2)) ///
>         order(*lntrade* educ*)

------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
                              (1)             (2)             (3)             (4)         
>     (5)             (6)             (7)             (8)             (9)   
                     protest_ikd2    protest_ikd2    protest_ikd2    protest_ikd2    prote
> st_ikd2    protest_ikd2    protest_ikd2    protest_ikd2    protest_ikd2   
                             b/se            b/se            b/se            b/se         
>    b/se            b/se            b/se            b/se            b/se   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
protest_ikd2                                                                              
>                                                                           
L.Imports+Exports ~n        0.415*          0.379           0.585**         0.376         
>   0.532**         0.459*          0.604**         0.564**         0.551** 
                           (0.24)          (0.25)          (0.24)          (0.25)         
>  (0.24)          (0.24)          (0.24)          (0.24)          (0.24)   
L.Imports+Exports ~#       -0.041**        -0.042**        -0.051***       -0.043**       
>  -0.049***       -0.040**        -0.058***       -0.048***       -0.048***
                           (0.02)          (0.02)          (0.02)          (0.02)         
>  (0.02)          (0.02)          (0.02)          (0.02)          (0.02)   
educshare                   0.082           0.070           0.081           0.079         
>   0.074           0.059           0.104*          0.063           0.086   
                           (0.06)          (0.06)          (0.06)          (0.06)         
>  (0.06)          (0.06)          (0.06)          (0.06)          (0.06)   
educshare                   0.000                                                         
>                                                                           
                              (.)                                                         
>                                                                           
L.reg_pop                   0.474***        0.409***        0.477***        0.386***      
>   0.407***        0.512***        0.546***        0.452***        0.429***
                           (0.07)          (0.07)          (0.07)          (0.08)         
>  (0.08)          (0.08)          (0.07)          (0.07)          (0.07)   
L.reg_urbanshare            0.017           0.022**         0.019*          0.021**       
>   0.018*          0.018*          0.013           0.017*          0.014   
                           (0.01)          (0.01)          (0.01)          (0.01)         
>  (0.01)          (0.01)          (0.01)          (0.01)          (0.01)   
L.lngrppc                  -0.644***        0.006          -0.410***        0.335         
>  -0.663***       -0.579***       -0.839***       -0.831***       -0.541***
                           (0.13)          (0.24)          (0.16)          (0.27)         
>  (0.14)          (0.13)          (0.14)          (0.18)          (0.13)   
L.reg_grpgr                -0.416**        -0.659***       -0.489**        -0.755***      
>  -0.466**        -0.501**        -0.269          -0.456**        -0.470** 
                           (0.21)          (0.21)          (0.20)          (0.21)         
>  (0.20)          (0.20)          (0.20)          (0.21)          (0.21)   
L.reg_levelofunempl        -0.028          -0.013          -0.042          -0.005         
>  -0.032          -0.039          -0.042          -0.040          -0.041   
                           (0.03)          (0.03)          (0.03)          (0.03)         
>  (0.03)          (0.03)          (0.03)          (0.03)          (0.03)   
L.lnnews                   -0.048          -0.012          -0.067          -0.014         
>  -0.035          -0.016          -0.069          -0.079          -0.035   
                           (0.11)          (0.11)          (0.11)          (0.11)         
>  (0.11)          (0.11)          (0.11)          (0.11)          (0.11)   
L.lndistmos                 0.185**         0.208***        0.214***        0.215***      
>   0.190**         0.171**         0.236***        0.179**         0.155** 
                           (0.08)          (0.08)          (0.08)          (0.08)         
>  (0.08)          (0.08)          (0.08)          (0.08)          (0.08)   
L.lnroadden                -0.146          -0.125          -0.050          -0.061         
>  -0.108          -0.137          -0.250**        -0.252**        -0.333** 
                           (0.10)          (0.10)          (0.11)          (0.10)         
>  (0.10)          (0.10)          (0.10)          (0.12)          (0.13)   
reg_density                 0.061          -0.070           0.055          -0.078         
>  -0.007          -0.009          -0.024           0.040           0.029   
                           (0.08)          (0.12)          (0.07)          (0.11)         
>  (0.10)          (0.09)          (0.11)          (0.09)          (0.09)   
L.FDI inflows (% o~n        0.709***        0.108           0.075           0.086         
>   0.118           0.123           0.116           0.140           0.092   
                           (0.23)          (0.10)          (0.10)          (0.10)         
>  (0.10)          (0.10)          (0.10)          (0.10)          (0.10)   
oL.FDI inflows (% ~n        0.000                                                         
>                                                                           
                              (.)                                                         
>                                                                           
L.FDI inflows (% o~u       -0.051***                                                      
>                                                                           
                           (0.02)                                                         
>                                                                           
reg_costliving                             -0.000***                                      
>                                                                           
                                           (0.00)                                         
>                                                                           
reg_belowcost                                               0.043**                       
>                                                                           
                                                           (0.02)                         
>                                                                           
reg_gdsfixed                                                               -0.000***      
>                                                                           
                                                                           (0.00)         
>                                                                           
reg_mincgini                                                                              
>   5.387                                                                   
                                                                                          
>  (4.30)                                                                   
reg_interregmigrat~n                                                                      
>                  -0.000*                                                  
                                                                                          
>                  (0.00)                                                   
reg_infantmort                                                                            
>                                  -0.140***                                
                                                                                          
>                                  (0.03)                                   
reg_lifeexp                                                                               
>                                                   0.086*                  
                                                                                          
>                                                  (0.05)                   
reg_ibr                                                                                   
>                                                                  -0.397** 
                                                                                          
>                                                                  (0.18)   
Constant                    5.361**        -1.597           1.248          -5.229         
>   3.296           4.443**         9.097***        2.554           5.731***
                           (2.15)          (2.96)          (2.69)          (3.27)         
>  (2.42)          (2.12)          (2.39)          (2.40)          (2.16)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
ln_r                                                                                      
>                                                                           
Constant                    0.366*          0.369**         0.382**         0.367**       
>   0.375**         0.407**         0.410**         0.424**         0.431** 
                           (0.19)          (0.19)          (0.19)          (0.18)         
>  (0.19)          (0.19)          (0.19)          (0.19)          (0.19)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
ln_s                                                                                      
>                                                                           
Constant                    1.166***        1.171***        1.191***        1.144***      
>   1.193***        1.245***        1.259***        1.295***        1.302***
                           (0.27)          (0.26)          (0.28)          (0.25)         
>  (0.27)          (0.27)          (0.27)          (0.27)          (0.28)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
# of observations             427             427             427             427         
>     427             427             427             425             427   
# of regions                   75              75              75              75         
>      75              75              75              75              75   
Prob > Chi2                 0.000           0.000           0.000           0.000         
>   0.000           0.000           0.000           0.000           0.000   
Log likelihood           -1245.42        -1244.66        -1247.13        -1241.78        -
> 1248.59        -1247.21        -1240.65        -1241.06        -1246.69   
AIC                       2524.83         2523.32         2528.27         2517.57         
> 2531.18         2528.42         2515.29         2516.12         2527.37   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------

. 
. *** Additional robustness checks
. * Longer average education
. xtnbreg protest_ikd2 $X4 $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -74606.474  (not concave)
Iteration 1:   log likelihood = -62683.468  
Iteration 2:   log likelihood =  -47710.12  (backed up)
Iteration 3:   log likelihood = -40926.332  (backed up)
Iteration 4:   log likelihood = -23462.547  
Iteration 5:   log likelihood = -6543.2598  
Iteration 6:   log likelihood = -2690.3042  
Iteration 7:   log likelihood = -2594.0904  
Iteration 8:   log likelihood = -2593.0863  
Iteration 9:   log likelihood = -2593.0861  
Iteration 10:  log likelihood = -2593.0861  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1390.6446  (not concave)
Iteration 2:   log likelihood = -1369.1213  
Iteration 3:   log likelihood = -1286.4542  
Iteration 4:   log likelihood = -1282.0091  
Iteration 5:   log likelihood = -1281.9762  
Iteration 6:   log likelihood = -1281.9762  

Fitting full model:

Iteration 0:   log likelihood = -1396.5537  (not concave)
Iteration 1:   log likelihood = -1359.8523  (not concave)
Iteration 2:   log likelihood = -1316.1719  
Iteration 3:   log likelihood = -1264.9636  
Iteration 4:   log likelihood = -1228.0013  
Iteration 5:   log likelihood = -1221.7897  
Iteration 6:   log likelihood = -1221.6549  
Iteration 7:   log likelihood = -1221.6545  
Iteration 8:   log likelihood = -1221.6545  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      82.31
Log likelihood  = -1221.6545                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
           protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                lntrade |
                    L1. |   .2987979   .2084062     1.43   0.152    -.1096707    .7072666
                        |
             educshare2 |   .0507566   .0610593     0.83   0.406    -.0689174    .1704305
                        |
cL.lntrade#c.educshare2 |  -.0379664   .0185725    -2.04   0.041    -.0743678   -.0015649
                        |
                reg_pop |
                    L1. |    .447772   .0735509     6.09   0.000     .3036148    .5919291
                        |
         reg_urbanshare |
                    L1. |   .0110059   .0111882     0.98   0.325    -.0109226    .0329344
                        |
                lngrppc |
                    L1. |  -.6161067   .1430444    -4.31   0.000    -.8964686   -.3357448
                        |
              reg_grpgr |
                    L1. |  -.4567792   .2101368    -2.17   0.030    -.8686396   -.0449187
                        |
      reg_levelofunempl |
                    L1. |  -.0353253   .0300278    -1.18   0.239    -.0941787    .0235282
                        |
                 lnnews |
                    L1. |  -.0595012   .1150958    -0.52   0.605    -.2850848    .1660825
                        |
              lndistmos |
                    L1. |    .187277   .0788669     2.37   0.018     .0327007    .3418532
                        |
              lnroadden |
                    L1. |  -.0598191   .1114382    -0.54   0.591     -.278234    .1585957
                        |
            reg_density |   .0606624   .0770073     0.79   0.431    -.0902692     .211594
                        |
                  lnfdi |
                    L1. |   .0654045   .0998643     0.65   0.513     -.130326    .2611349
                        |
                  rents |    .011367   .0094056     1.21   0.227    -.0070677    .0298016
              pressfill |   .0510777   .0921366     0.55   0.579    -.1295067     .231662
       KPRFmandateshare |  -.0266063   .0111014    -2.40   0.017    -.0483646    -.004848
             pctRussian |   .0057815   .0061242     0.94   0.345    -.0062217    .0177848
                  _cons |   5.312918   2.277907     2.33   0.020      .848301    9.777534
------------------------+----------------------------------------------------------------
                  /ln_r |   .3774688   .1937479                       -.00227    .7572077
                  /ln_s |   1.208298   .2747329                      .6698314    1.746765
------------------------+----------------------------------------------------------------
                      r |   1.458588   .2825983                      .9977326    2.132314
                      s |   3.347782   .9197458                      1.953908    5.736014
-----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 120.64               Prob >= chibar2 = 0.000

. est store robust1

. * Tertiary education only
. xtnbreg protest_ikd2 cl.lntrade##c.educshareter $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -74803.822  (not concave)
Iteration 1:   log likelihood = -62904.819  
Iteration 2:   log likelihood = -36173.469  (backed up)
Iteration 3:   log likelihood = -15918.922  (backed up)
Iteration 4:   log likelihood = -9205.3617  
Iteration 5:   log likelihood = -5353.4944  
Iteration 6:   log likelihood = -2897.2793  
Iteration 7:   log likelihood = -2692.9877  
Iteration 8:   log likelihood =  -2682.344  
Iteration 9:   log likelihood = -2682.3194  
Iteration 10:  log likelihood = -2682.3194  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1391.1425  (not concave)
Iteration 2:   log likelihood =  -1370.261  
Iteration 3:   log likelihood = -1293.9577  
Iteration 4:   log likelihood = -1287.2681  
Iteration 5:   log likelihood = -1287.1097  
Iteration 6:   log likelihood = -1287.1094  
Iteration 7:   log likelihood = -1287.1094  

Fitting full model:

Iteration 0:   log likelihood = -1397.2021  (not concave)
Iteration 1:   log likelihood = -1282.5254  
Iteration 2:   log likelihood = -1243.4275  (not concave)
Iteration 3:   log likelihood =  -1218.766  
Iteration 4:   log likelihood = -1213.8091  
Iteration 5:   log likelihood = -1213.7199  
Iteration 6:   log likelihood = -1213.7197  
Iteration 7:   log likelihood = -1213.7197  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      95.84
Log likelihood  = -1213.7197                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------
     protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
          lntrade |
              L1. |   .9080983   .2489603     3.65   0.000     .4201452    1.396051
                  |
     educshareter |   .3055361     .10531     2.90   0.004     .0991322    .5119399
                  |
       cL.lntrade#|
   c.educshareter |  -.1260907   .0324865    -3.88   0.000    -.1897631   -.0624182
                  |
          reg_pop |
              L1. |   .5354462   .0803357     6.67   0.000     .3779911    .6929014
                  |
   reg_urbanshare |
              L1. |   .0234313   .0118275     1.98   0.048     .0002499    .0466128
                  |
          lngrppc |
              L1. |  -.6138881   .1408656    -4.36   0.000    -.8899796   -.3377966
                  |
        reg_grpgr |
              L1. |   -.466028   .2016022    -2.31   0.021    -.8611611   -.0708949
                  |
reg_levelofunempl |
              L1. |  -.0525498   .0295384    -1.78   0.075    -.1104441    .0053444
                  |
           lnnews |
              L1. |  -.0233146    .110156    -0.21   0.832    -.2392165    .1925872
                  |
        lndistmos |
              L1. |   .2005799   .0802289     2.50   0.012      .043334    .3578257
                  |
        lnroadden |
              L1. |  -.1368478   .1157693    -1.18   0.237    -.3637516    .0900559
                  |
      reg_density |   .3024954   .1251323     2.42   0.016     .0572406    .5477503
                  |
            lnfdi |
              L1. |   .1074832    .100953     1.06   0.287    -.0903809    .3053474
                  |
            rents |   .0048891   .0101354     0.48   0.630    -.0149759     .024754
        pressfill |   .0429501   .0926323     0.46   0.643    -.1386058    .2245061
 KPRFmandateshare |   -.022678   .0112381    -2.02   0.044    -.0447042   -.0006518
       pctRussian |   .0062967   .0062488     1.01   0.314    -.0059506     .018544
            _cons |   2.558911   2.266354     1.13   0.259    -1.883061    7.000883
------------------+----------------------------------------------------------------
            /ln_r |     .32735   .1808107                     -.0270324    .6817325
            /ln_s |   1.040172   .2516964                       .546856    1.533488
------------------+----------------------------------------------------------------
                r |   1.387287   .2508363                      .9733297      1.9773
                s |   2.829703   .7122262                      1.727812    4.634312
-----------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 146.78               Prob >= chibar2 = 0.000

. est store robust2

. * Secondary + 2*tertiary education
. xtnbreg protest_ikd2 cl.lntrade##c.educshareind $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -75323.684  (not concave)
Iteration 1:   log likelihood = -63283.879  
Iteration 2:   log likelihood = -34387.578  (backed up)
Iteration 3:   log likelihood = -18698.109  
Iteration 4:   log likelihood = -10253.735  
Iteration 5:   log likelihood = -4229.4803  
Iteration 6:   log likelihood = -2700.6689  
Iteration 7:   log likelihood = -2651.3754  
Iteration 8:   log likelihood = -2650.1567  
Iteration 9:   log likelihood = -2650.1557  
Iteration 10:  log likelihood = -2650.1557  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1390.7492  (not concave)
Iteration 2:   log likelihood = -1369.4929  
Iteration 3:   log likelihood = -1288.2948  
Iteration 4:   log likelihood = -1283.2551  
Iteration 5:   log likelihood = -1283.1933  
Iteration 6:   log likelihood = -1283.1932  

Fitting full model:

Iteration 0:   log likelihood = -1391.6788  (not concave)
Iteration 1:   log likelihood = -1296.8483  
Iteration 2:   log likelihood = -1264.3001  
Iteration 3:   log likelihood = -1223.2048  
Iteration 4:   log likelihood = -1213.7063  
Iteration 5:   log likelihood = -1213.2242  
Iteration 6:   log likelihood = -1213.2194  
Iteration 7:   log likelihood = -1213.2194  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =     102.24
Log likelihood  = -1213.2194                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------
     protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
          lntrade |
              L1. |   .7360651   .2394981     3.07   0.002     .2666575    1.205473
                  |
     educshareind |   .0963539   .0428654     2.25   0.025     .0123392    .1803686
                  |
       cL.lntrade#|
   c.educshareind |  -.0468843   .0128633    -3.64   0.000    -.0720959   -.0216728
                  |
          reg_pop |
              L1. |   .5484325   .0780312     7.03   0.000     .3954942    .7013707
                  |
   reg_urbanshare |
              L1. |   .0216901    .011607     1.87   0.062    -.0010591    .0444394
                  |
          lngrppc |
              L1. |    -.60679   .1409897    -4.30   0.000    -.8831247   -.3304552
                  |
        reg_grpgr |
              L1. |  -.4595109   .2007957    -2.29   0.022    -.8530632   -.0659586
                  |
reg_levelofunempl |
              L1. |  -.0465992   .0290881    -1.60   0.109    -.1036109    .0104126
                  |
           lnnews |
              L1. |  -.0306965   .1116809    -0.27   0.783    -.2495871     .188194
                  |
        lndistmos |
              L1. |   .1893234   .0786722     2.41   0.016     .0351288     .343518
                  |
        lnroadden |
              L1. |  -.1128809    .112104    -1.01   0.314    -.3326007    .1068388
                  |
      reg_density |   .1545482   .0778506     1.99   0.047     .0019638    .3071326
                  |
            lnfdi |
              L1. |   .0879534   .0975805     0.90   0.367    -.1033008    .2792076
                  |
            rents |   .0060647   .0097178     0.62   0.533    -.0129818    .0251112
        pressfill |   .0504577   .0897702     0.56   0.574    -.1254887     .226404
 KPRFmandateshare |  -.0245372   .0109868    -2.23   0.026    -.0460709   -.0030034
       pctRussian |    .005818   .0062222     0.94   0.350    -.0063772    .0180132
            _cons |   3.241054   2.312392     1.40   0.161     -1.29115    7.773259
------------------+----------------------------------------------------------------
            /ln_r |   .3871128   .1862063                      .0221551    .7520705
            /ln_s |     1.1679   .2558739                      .6663959    1.669403
------------------+----------------------------------------------------------------
                r |   1.472723   .2742303                      1.022402    2.121388
                s |   3.215232   .8226942                      1.947207    5.308999
-----------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 139.95               Prob >= chibar2 = 0.000

. est store robust3

. * Exports only
. xtnbreg protest_ikd2 cl.lnexport##c.educshare $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -73960.948  (not concave)
Iteration 1:   log likelihood = -62215.556  
Iteration 2:   log likelihood = -27403.828  (backed up)
Iteration 3:   log likelihood = -10336.294  (backed up)
Iteration 4:   log likelihood = -5246.8291  
Iteration 5:   log likelihood = -2998.0132  
Iteration 6:   log likelihood =  -2646.928  
Iteration 7:   log likelihood =  -2636.094  
Iteration 8:   log likelihood =  -2636.005  
Iteration 9:   log likelihood =  -2636.005  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1390.7002  (not concave)
Iteration 2:   log likelihood = -1369.3402  
Iteration 3:   log likelihood =  -1289.298  
Iteration 4:   log likelihood = -1283.2416  
Iteration 5:   log likelihood = -1283.1507  
Iteration 6:   log likelihood = -1283.1506  

Fitting full model:

Iteration 0:   log likelihood = -1387.3909  (not concave)
Iteration 1:   log likelihood = -1291.5416  
Iteration 2:   log likelihood = -1245.0369  (not concave)
Iteration 3:   log likelihood = -1228.2513  
Iteration 4:   log likelihood = -1220.6728  
Iteration 5:   log likelihood = -1220.4386  
Iteration 6:   log likelihood = -1220.4353  
Iteration 7:   log likelihood = -1220.4353  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      85.41
Log likelihood  = -1220.4353                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
           protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
               lnexport |
                    L1. |   .1796708   .1997421     0.90   0.368    -.2118166    .5711582
                        |
              educshare |  -.0159438   .0396283    -0.40   0.687    -.0936139    .0617263
                        |
cL.lnexport#c.educshare |  -.0228372   .0132319    -1.73   0.084    -.0487712    .0030967
                        |
                reg_pop |
                    L1. |   .4722824   .0741467     6.37   0.000     .3269576    .6176073
                        |
         reg_urbanshare |
                    L1. |   .0095102   .0108804     0.87   0.382    -.0118151    .0308355
                        |
                lngrppc |
                    L1. |  -.6514993   .1430954    -4.55   0.000    -.9319613   -.3710374
                        |
              reg_grpgr |
                    L1. |   -.414774   .2106135    -1.97   0.049    -.8275688   -.0019791
                        |
      reg_levelofunempl |
                    L1. |  -.0369458   .0294945    -1.25   0.210    -.0947539    .0208623
                        |
                 lnnews |
                    L1. |  -.1045464    .113401    -0.92   0.357    -.3268083    .1177155
                        |
              lndistmos |
                    L1. |    .185259   .0817221     2.27   0.023     .0250865    .3454314
                        |
              lnroadden |
                    L1. |  -.1039334   .1111519    -0.94   0.350    -.3217872    .1139203
                        |
            reg_density |   .0539533   .0772715     0.70   0.485     -.097496    .2054026
                        |
                  lnfdi |
                    L1. |   .0822702   .0989617     0.83   0.406    -.1116911    .2762315
                        |
                  rents |   .0122145    .009725     1.26   0.209    -.0068461    .0312751
              pressfill |   .0577395   .0912737     0.63   0.527    -.1211536    .2366326
       KPRFmandateshare |  -.0257476   .0110346    -2.33   0.020     -.047375   -.0041203
             pctRussian |   .0065157    .006363     1.02   0.306    -.0059555    .0189869
                  _cons |    7.01244   2.139419     3.28   0.001     2.819255    11.20563
------------------------+----------------------------------------------------------------
                  /ln_r |   .3600449    .190605                     -.0135339    .7336238
                  /ln_s |   1.147859   .2670622                       .624427    1.671292
------------------------+----------------------------------------------------------------
                      r |   1.433394    .273212                      .9865572    2.082614
                      s |    3.15144   .8416305                      1.867176    5.319034
-----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 125.43               Prob >= chibar2 = 0.000

. est store robust4

. * Imports only
. xtnbreg protest_ikd2 cl.lnimport##c.educshare $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -73990.586  (not concave)
Iteration 1:   log likelihood = -62189.561  
Iteration 2:   log likelihood = -36406.756  (backed up)
Iteration 3:   log likelihood = -14059.286  (backed up)
Iteration 4:   log likelihood = -7719.2755  
Iteration 5:   log likelihood = -4287.9305  
Iteration 6:   log likelihood = -2730.3405  
Iteration 7:   log likelihood = -2635.0152  
Iteration 8:   log likelihood = -2633.3693  
Iteration 9:   log likelihood = -2633.3682  
Iteration 10:  log likelihood = -2633.3682  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1390.7525  (not concave)
Iteration 2:   log likelihood = -1369.4622  
Iteration 3:   log likelihood = -1289.0575  
Iteration 4:   log likelihood = -1282.2972  
Iteration 5:   log likelihood = -1282.1968  
Iteration 6:   log likelihood = -1282.1967  

Fitting full model:

Iteration 0:   log likelihood = -1379.3739  (not concave)
Iteration 1:   log likelihood = -1309.3385  (not concave)
Iteration 2:   log likelihood = -1295.1224  (not concave)
Iteration 3:   log likelihood = -1255.8267  
Iteration 4:   log likelihood = -1247.5107  
Iteration 5:   log likelihood =  -1226.111  
Iteration 6:   log likelihood = -1222.3888  
Iteration 7:   log likelihood =   -1221.92  
Iteration 8:   log likelihood = -1221.9147  
Iteration 9:   log likelihood = -1221.9147  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      80.76
Log likelihood  = -1221.9147                    Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
           protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
               lnimport |
                    L1. |   .0178627   .1692086     0.11   0.916    -.3137801    .3495055
                        |
              educshare |  -.0773438   .0279554    -2.77   0.006    -.1321354   -.0225523
                        |
cL.lnimport#c.educshare |   .0059636   .0114374     0.52   0.602    -.0164533    .0283804
                        |
                reg_pop |
                    L1. |   .4263631   .0725097     5.88   0.000     .2842467    .5684795
                        |
         reg_urbanshare |
                    L1. |   .0026348   .0106828     0.25   0.805    -.0183032    .0235727
                        |
                lngrppc |
                    L1. |  -.6577101   .1461634    -4.50   0.000    -.9441851   -.3712352
                        |
              reg_grpgr |
                    L1. |  -.4366632   .2139881    -2.04   0.041    -.8560721   -.0172543
                        |
      reg_levelofunempl |
                    L1. |  -.0253703   .0293068    -0.87   0.387    -.0828106    .0320699
                        |
                 lnnews |
                    L1. |  -.1297194   .1122866    -1.16   0.248    -.3497972    .0903584
                        |
              lndistmos |
                    L1. |   .1746786   .0824096     2.12   0.034     .0131588    .3361984
                        |
              lnroadden |
                    L1. |   -.138087   .1168454    -1.18   0.237    -.3670998    .0909258
                        |
            reg_density |   .0265699   .0963836     0.28   0.783    -.1623385    .2154782
                        |
                  lnfdi |
                    L1. |   .0662468   .1017437     0.65   0.515    -.1331672    .2656608
                        |
                  rents |   .0120045   .0097335     1.23   0.217    -.0070729    .0310818
              pressfill |   .0700211    .094822     0.74   0.460    -.1158267    .2558689
       KPRFmandateshare |  -.0263621   .0112595    -2.34   0.019    -.0484303   -.0042939
             pctRussian |   .0031906   .0062469     0.51   0.610    -.0090532    .0154343
                  _cons |   8.580237   2.174536     3.95   0.000     4.318225    12.84225
------------------------+----------------------------------------------------------------
                  /ln_r |   .3370506   .1900366                     -.0354143    .7095156
                  /ln_s |   1.118016   .2678116                      .5931145    1.642917
------------------------+----------------------------------------------------------------
                      r |    1.40081   .2662052                      .9652054    2.033006
                      s |   3.058778   .8191761                      1.809616    5.170227
-----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 120.56               Prob >= chibar2 = 0.000

. est store robust5

. * Fixed effects negative binomial (without secondary education share)
. xtnbreg protest_ikd2 cl.lntrade cl.lntrade#c.educshare $C1 if sample1==1, fe
note: 3 groups (8 obs) dropped because of all zero outcomes

Iteration 0:   log likelihood = -2092.8223  
Iteration 1:   log likelihood = -1492.1643  (not concave)
Iteration 2:   log likelihood = -989.39803  
Iteration 3:   log likelihood = -902.63038  (not concave)
Iteration 4:   log likelihood = -878.98251  
Iteration 5:   log likelihood = -876.90421  
Iteration 6:   log likelihood = -876.74736  
Iteration 7:   log likelihood = -876.74625  
Iteration 8:   log likelihood = -876.74625  

Conditional FE negative binomial regression     Number of obs     =        419
Group variable: rcode                           Number of groups  =         72

                                                Obs per group:
                                                              min =          3
                                                              avg =        5.8
                                                              max =          6

                                                Wald chi2(12)     =      50.51
Log likelihood  = -876.74625                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .1894912   .1255721     1.51   0.131    -.0566255     .435608
                       |
cL.lntrade#c.educshare |  -.0274392   .0074993    -3.66   0.000    -.0421375    -.012741
                       |
               reg_pop |
                   L1. |   .2847201   .0950741     2.99   0.003     .0983782    .4710619
                       |
        reg_urbanshare |
                   L1. |   .0021032   .0139622     0.15   0.880    -.0252623    .0294687
                       |
               lngrppc |
                   L1. |  -.6442807   .1342035    -4.80   0.000    -.9073148   -.3812467
                       |
             reg_grpgr |
                   L1. |  -.4413274   .2151718    -2.05   0.040    -.8630563   -.0195985
                       |
     reg_levelofunempl |
                   L1. |  -.0280085   .0314637    -0.89   0.373    -.0896763    .0336593
                       |
                lnnews |
                   L1. |  -.0315678    .120961    -0.26   0.794     -.268647    .2055114
                       |
             lndistmos |
                   L1. |   .1067674   .0942028     1.13   0.257    -.0778667    .2914015
                       |
             lnroadden |
                   L1. |  -.2978271   .1465561    -2.03   0.042    -.5850718   -.0105824
                       |
           reg_density |   .0700122   .1144266     0.61   0.541    -.1542598    .2942841
                       |
                 lnfdi |
                   L1. |   .0723859   .1082691     0.67   0.504    -.1398177    .2845894
                       |
                 _cons |   9.307993   2.183154     4.26   0.000     5.029089     13.5869
----------------------------------------------------------------------------------------

. est store robust6

. * Federal district fixed effects
. nbreg protest_ikd2 $X2 $C1 $C2 i.dname if sample1==1

Fitting Poisson model:

Iteration 0:   log likelihood = -75221.696  (not concave)
Iteration 1:   log likelihood = -51153.756  
Iteration 2:   log likelihood = -32727.217  (backed up)
Iteration 3:   log likelihood = -29517.838  (backed up)
Iteration 4:   log likelihood = -15402.192  (backed up)
Iteration 5:   log likelihood = -14274.269  
Iteration 6:   log likelihood = -8227.4025  
Iteration 7:   log likelihood =  -2897.861  
Iteration 8:   log likelihood = -2433.3934  
Iteration 9:   log likelihood = -2397.9251  
Iteration 10:  log likelihood = -2397.7938  
Iteration 11:  log likelihood = -2397.7938  

Fitting constant-only model:

Iteration 0:   log likelihood = -1540.3992  
Iteration 1:   log likelihood = -1432.0601  
Iteration 2:   log likelihood = -1431.1607  
Iteration 3:   log likelihood = -1431.1601  
Iteration 4:   log likelihood = -1431.1601  

Fitting full model:

Iteration 0:   log likelihood = -1351.6265  
Iteration 1:   log likelihood = -1275.2554  
Iteration 2:   log likelihood = -1262.7858  
Iteration 3:   log likelihood = -1262.3679  
Iteration 4:   log likelihood =  -1262.367  
Iteration 5:   log likelihood =  -1262.367  

Negative binomial regression                    Number of obs     =        416
                                                LR chi2(24)       =     337.59
Dispersion     = mean                           Prob > chi2       =     0.0000
Log likelihood =  -1262.367                     Pseudo R2         =     0.1179

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |    .752452   .2474167     3.04   0.002     .2675242     1.23738
                       |
             educshare |   .1674871   .0557432     3.00   0.003     .0582324    .2767417
                       |
cL.lntrade#c.educshare |  -.0596965   .0181598    -3.29   0.001    -.0952892   -.0241039
                       |
               reg_pop |
                   L1. |    .665872   .0807047     8.25   0.000     .5076937    .8240503
                       |
        reg_urbanshare |
                   L1. |   .0322617   .0091315     3.53   0.000     .0143643     .050159
                       |
               lngrppc |
                   L1. |  -.2685186   .2056134    -1.31   0.192    -.6715135    .1344762
                       |
             reg_grpgr |
                   L1. |  -.5339794   .3031652    -1.76   0.078    -1.128172    .0602136
                       |
     reg_levelofunempl |
                   L1. |  -.0655132    .032437    -2.02   0.043    -.1290885   -.0019379
                       |
                lnnews |
                   L1. |  -.2696338   .1179523    -2.29   0.022     -.500816   -.0384516
                       |
             lndistmos |
                   L1. |   .3013856   .1004713     3.00   0.003     .1044654    .4983057
                       |
             lnroadden |
                   L1. |   .1655928   .0950556     1.74   0.081    -.0207129    .3518984
                       |
           reg_density |   .0305503   .0704699     0.43   0.665    -.1075681    .1686688
                       |
                 lnfdi |
                   L1. |   .1814713   .1229036     1.48   0.140    -.0594152    .4223579
                       |
                 rents |  -.0001332   .0079065    -0.02   0.987    -.0156297    .0153633
             pressfill |   .1517433   .0981386     1.55   0.122    -.0406048    .3440914
      KPRFmandateshare |  -.0220784   .0122395    -1.80   0.071    -.0460674    .0019106
            pctRussian |    .010017   .0050781     1.97   0.049     .0000642    .0199698
                       |
                 dname |
             Far East  |  -.2567884   .4385192    -0.59   0.558     -1.11627    .6026933
       North Caucasus  |  -.1544807   .4762472    -0.32   0.746    -1.087908    .7789467
           North West  |   .9348917   .2683041     3.48   0.000     .4090254    1.460758
              Siberia  |   .4279406   .3591174     1.19   0.233    -.2759166    1.131798
                South  |   .3762037   .3302091     1.14   0.255    -.2709942    1.023402
                 Ural  |  -.1016121   .3804563    -0.27   0.789    -.8472929    .6440686
                Volga  |   .4829396    .249548     1.94   0.053    -.0061655    .9720447
                       |
                 _cons |  -2.370843   2.675944    -0.89   0.376    -7.615597    2.873911
-----------------------+----------------------------------------------------------------
              /lnalpha |   .0322531   .0856473                     -.1356124    .2001187
-----------------------+----------------------------------------------------------------
                 alpha |   1.032779   .0884547                       .873181    1.221548
----------------------------------------------------------------------------------------
LR test of alpha=0: chibar2(01) = 2270.85              Prob >= chibar2 = 0.000

. est store robust7

. * Economic region fixed effects
. nbreg protest_ikd2 $X2 $C1 $C2 i.ename if sample1==1

Fitting Poisson model:

Iteration 0:   log likelihood = -75777.759  (not concave)
Iteration 1:   log likelihood =  -51529.36  
Iteration 2:   log likelihood = -40966.781  (backed up)
Iteration 3:   log likelihood = -25413.118  (backed up)
Iteration 4:   log likelihood = -16546.377  (backed up)
Iteration 5:   log likelihood = -12794.408  
Iteration 6:   log likelihood = -4525.8691  
Iteration 7:   log likelihood =  -2424.413  
Iteration 8:   log likelihood = -2368.8102  
Iteration 9:   log likelihood = -2368.6559  
Iteration 10:  log likelihood = -2368.6559  

Fitting constant-only model:

Iteration 0:   log likelihood = -1540.3992  
Iteration 1:   log likelihood = -1432.0601  
Iteration 2:   log likelihood = -1431.1607  
Iteration 3:   log likelihood = -1431.1601  
Iteration 4:   log likelihood = -1431.1601  

Fitting full model:

Iteration 0:   log likelihood =  -1350.674  
Iteration 1:   log likelihood =  -1272.227  
Iteration 2:   log likelihood = -1252.9552  
Iteration 3:   log likelihood = -1251.7188  
Iteration 4:   log likelihood =  -1251.716  
Iteration 5:   log likelihood =  -1251.716  

Negative binomial regression                    Number of obs     =        416
                                                LR chi2(28)       =     358.89
Dispersion     = mean                           Prob > chi2       =     0.0000
Log likelihood =  -1251.716                     Pseudo R2         =     0.1254

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .6759652   .2428829     2.78   0.005     .1999234    1.152007
                       |
             educshare |   .1767218   .0541189     3.27   0.001     .0706506    .2827929
                       |
cL.lntrade#c.educshare |  -.0571095   .0174983    -3.26   0.001    -.0914055   -.0228135
                       |
               reg_pop |
                   L1. |   .7703032   .0953851     8.08   0.000      .583352    .9572545
                       |
        reg_urbanshare |
                   L1. |   .0205015   .0100914     2.03   0.042     .0007227    .0402802
                       |
               lngrppc |
                   L1. |  -.3091066   .2042458    -1.51   0.130     -.709421    .0912077
                       |
             reg_grpgr |
                   L1. |  -.5475426   .3002801    -1.82   0.068    -1.136081    .0409956
                       |
     reg_levelofunempl |
                   L1. |   -.088795   .0346294    -2.56   0.010    -.1566674   -.0209226
                       |
                lnnews |
                   L1. |   -.283777   .1286868    -2.21   0.027    -.5359986   -.0315554
                       |
             lndistmos |
                   L1. |   .4053288    .120797     3.36   0.001      .168571    .6420866
                       |
             lnroadden |
                   L1. |  -.0569416   .1205113    -0.47   0.637    -.2931394    .1792563
                       |
           reg_density |   .0044907   .0672866     0.07   0.947    -.1273886    .1363701
                       |
                 lnfdi |
                   L1. |    .332364   .1243815     2.67   0.008     .0885807    .5761473
                       |
                 rents |   -.010511   .0084459    -1.24   0.213    -.0270648    .0060427
             pressfill |   .1602248   .0974864     1.64   0.100     -.030845    .3512947
      KPRFmandateshare |  -.0206091   .0119723    -1.72   0.085    -.0440744    .0028563
            pctRussian |    .004961   .0051056     0.97   0.331    -.0050457    .0149677
                       |
                 ename |
  Central Black Earth  |  -.3917671   .3585168    -1.09   0.275    -1.094447     .310913
        East Siberian  |   -.757429   .5207402    -1.45   0.146    -1.778061    .2632032
          Far Eastern  |  -1.100962   .5353109    -2.06   0.040    -2.150152   -.0517722
          Kaliningrad  |   1.933983   .5542597     3.49   0.000     .8476536    3.020312
       North Caucasus  |  -1.045888   .4988171    -2.10   0.036    -2.023552   -.0682246
             Northern  |   .4697688   .3763474     1.25   0.212    -.2678585    1.207396
         Northwestern  |   .5767227   .3658634     1.58   0.115    -.1403563    1.293802
                 Ural  |   .2354725   .3992653     0.59   0.555    -.5470732    1.018018
                Volga  |   .2358188   .3300882     0.71   0.475    -.4111421    .8827798
         Volga-Vyatka  |   .0213588   .3297112     0.06   0.948    -.6248633     .667581
        West Siberian  |   .0154251   .4567487     0.03   0.973     -.879786    .9106361
                       |
                 _cons |   .0598307   2.774768     0.02   0.983    -5.378615    5.498277
-----------------------+----------------------------------------------------------------
              /lnalpha |  -.0234888   .0864057                     -.1928408    .1458632
-----------------------+----------------------------------------------------------------
                 alpha |   .9767849   .0843997                      .8246132    1.157038
----------------------------------------------------------------------------------------
LR test of alpha=0: chibar2(01) = 2233.88              Prob >= chibar2 = 0.000

. est store robust8

. * Small regions only (< median)
. gen smallregionm = 0

. replace smallregion = 1 if reg_area<=75
(672 real changes made)

. xtnbreg protest_ikd2 $X2 $C1 $C2 if sample1==1 & smallregion==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -25569.746  (not concave)
Iteration 1:   log likelihood =  -18767.42  
Iteration 2:   log likelihood = -14518.685  (backed up)
Iteration 3:   log likelihood = -9417.0216  (backed up)
Iteration 4:   log likelihood = -4634.3723  
Iteration 5:   log likelihood = -2801.5636  
Iteration 6:   log likelihood = -1239.6896  
Iteration 7:   log likelihood = -1134.3951  
Iteration 8:   log likelihood = -1131.7154  
Iteration 9:   log likelihood = -1131.7102  
Iteration 10:  log likelihood = -1131.7102  

Iteration 0:   log likelihood = -2725.1239  
Iteration 1:   log likelihood = -730.86013  
Iteration 2:   log likelihood = -695.40485  
Iteration 3:   log likelihood = -678.42574  
Iteration 4:   log likelihood = -678.40128  
Iteration 5:   log likelihood = -678.40128  

Iteration 0:   log likelihood = -678.40128  (not concave)
Iteration 1:   log likelihood = -644.35761  
Iteration 2:   log likelihood = -607.66419  
Iteration 3:   log likelihood = -589.86575  
Iteration 4:   log likelihood = -585.00992  
Iteration 5:   log likelihood = -584.94956  
Iteration 6:   log likelihood = -584.94948  
Iteration 7:   log likelihood = -584.94948  

Fitting full model:

Iteration 0:   log likelihood = -691.74967  (not concave)
Iteration 1:   log likelihood = -611.14717  
Iteration 2:   log likelihood = -599.53196  (not concave)
Iteration 3:   log likelihood = -570.08603  
Iteration 4:   log likelihood =   -564.776  
Iteration 5:   log likelihood = -561.92533  
Iteration 6:   log likelihood = -561.83934  
Iteration 7:   log likelihood = -561.83921  
Iteration 8:   log likelihood = -561.83921  

Random-effects negative binomial regression     Number of obs     =        209
Group variable: rcode                           Number of groups  =         37

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          3
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      62.49
Log likelihood  = -561.83921                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5006245   .4973356     1.01   0.314    -.4741354    1.475384
                       |
             educshare |   .1150385   .1061934     1.08   0.279    -.0930967    .3231737
                       |
cL.lntrade#c.educshare |  -.0553054   .0327365    -1.69   0.091    -.1194677    .0088568
                       |
               reg_pop |
                   L1. |   .3980582   .1706263     2.33   0.020     .0636367    .7324797
                       |
        reg_urbanshare |
                   L1. |   .0175726   .0211249     0.83   0.405    -.0238316    .0589767
                       |
               lngrppc |
                   L1. |  -.6872281    .248757    -2.76   0.006    -1.174783   -.1996733
                       |
             reg_grpgr |
                   L1. |   .1838418   .3329743     0.55   0.581    -.4687759    .8364596
                       |
     reg_levelofunempl |
                   L1. |   .0032758   .0480149     0.07   0.946    -.0908317    .0973832
                       |
                lnnews |
                   L1. |  -.3035748   .1635166    -1.86   0.063    -.6240614    .0169119
                       |
             lndistmos |
                   L1. |    .207935   .1570121     1.32   0.185    -.0998031     .515673
                       |
             lnroadden |
                   L1. |   .5699569   .3434942     1.66   0.097    -.1032795    1.243193
                       |
           reg_density |   .1696293   .0680051     2.49   0.013     .0363418    .3029169
                       |
                 lnfdi |
                   L1. |   .2768229   .1871471     1.48   0.139    -.0899788    .6436245
                       |
                 rents |   .0536482   .0223871     2.40   0.017     .0097703    .0975262
             pressfill |  -.1048484   .1657547    -0.63   0.527    -.4297217    .2200248
      KPRFmandateshare |  -.0553078   .0204594    -2.70   0.007    -.0954074   -.0152081
            pctRussian |   .0234798   .0089962     2.61   0.009     .0058476    .0411121
                 _cons |   1.888186   3.585764     0.53   0.598    -5.139782    8.916153
-----------------------+----------------------------------------------------------------
                 /ln_r |   .2768239   .2830678                     -.2779788    .8316267
                 /ln_s |    1.10847   .4315679                      .2626125    1.954328
-----------------------+----------------------------------------------------------------
                     r |   1.318934   .3733478                      .7573129    2.297052
                     s |   3.029719    1.30753                      1.300323     7.05917
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 46.22                Prob >= chibar2 = 0.000

. est store robust9

. * Table A5 in Online Appendix
. esttab robust1 robust2 robust3 robust4 robust5 robust6 robust7 robust8 robust9, ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g p ll aic, labels("# of observations" "# of regions" ///
>                 "Prob > Chi2" "Log likelihood" "AIC") fmt(0 0 3 2 2)) ///
>         order(*lntrade* educ*)

------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
                              (1)             (2)             (3)             (4)         
>     (5)             (6)             (7)             (8)             (9)   
                     protest_ikd2    protest_ikd2    protest_ikd2    protest_ikd2    prote
> st_ikd2    protest_ikd2    protest_ikd2    protest_ikd2    protest_ikd2   
                             b/se            b/se            b/se            b/se         
>    b/se            b/se            b/se            b/se            b/se   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
protest_ikd2                                                                              
>                                                                           
L.Imports+Exports ~n        0.299           0.908***        0.736***                      
>                   0.189           0.752***        0.676***        0.501   
                           (0.21)          (0.25)          (0.24)                         
>                  (0.13)          (0.25)          (0.24)          (0.50)   
L.Imports+Exports ~#       -0.038**                                                       
>                                                                           
                           (0.02)                                                         
>                                                                           
L.Imports+Exports ~#                       -0.126***                                      
>                                                                           
                                           (0.03)                                         
>                                                                           
L.Imports+Exports ~#                                       -0.047***                      
>                                                                           
                                                           (0.01)                         
>                                                                           
L.Imports+Exports ~#                                                                      
>                  -0.027***       -0.060***       -0.057***       -0.055*  
                                                                                          
>                  (0.01)          (0.02)          (0.02)          (0.03)   
educshare2                  0.051                                                         
>                                                                           
                           (0.06)                                                         
>                                                                           
educshareter                                0.306***                                      
>                                                                           
                                           (0.11)                                         
>                                                                           
educshareind                                                0.096**                       
>                                                                           
                                                           (0.04)                         
>                                                                           
educshare                                                                  -0.016         
>  -0.077***                        0.167***        0.177***        0.115   
                                                                           (0.04)         
>  (0.03)                          (0.06)          (0.05)          (0.11)   
L.reg_pop                   0.448***        0.535***        0.548***        0.472***      
>   0.426***        0.285***        0.666***        0.770***        0.398** 
                           (0.07)          (0.08)          (0.08)          (0.07)         
>  (0.07)          (0.10)          (0.08)          (0.10)          (0.17)   
L.reg_urbanshare            0.011           0.023**         0.022*          0.010         
>   0.003           0.002           0.032***        0.021**         0.018   
                           (0.01)          (0.01)          (0.01)          (0.01)         
>  (0.01)          (0.01)          (0.01)          (0.01)          (0.02)   
L.lngrppc                  -0.616***       -0.614***       -0.607***       -0.651***      
>  -0.658***       -0.644***       -0.269          -0.309          -0.687***
                           (0.14)          (0.14)          (0.14)          (0.14)         
>  (0.15)          (0.13)          (0.21)          (0.20)          (0.25)   
L.reg_grpgr                -0.457**        -0.466**        -0.460**        -0.415**       
>  -0.437**        -0.441**        -0.534*         -0.548*          0.184   
                           (0.21)          (0.20)          (0.20)          (0.21)         
>  (0.21)          (0.22)          (0.30)          (0.30)          (0.33)   
L.reg_levelofunempl        -0.035          -0.053*         -0.047          -0.037         
>  -0.025          -0.028          -0.066**        -0.089**         0.003   
                           (0.03)          (0.03)          (0.03)          (0.03)         
>  (0.03)          (0.03)          (0.03)          (0.03)          (0.05)   
L.lnnews                   -0.060          -0.023          -0.031          -0.105         
>  -0.130          -0.032          -0.270**        -0.284**        -0.304*  
                           (0.12)          (0.11)          (0.11)          (0.11)         
>  (0.11)          (0.12)          (0.12)          (0.13)          (0.16)   
L.lndistmos                 0.187**         0.201**         0.189**         0.185**       
>   0.175**         0.107           0.301***        0.405***        0.208   
                           (0.08)          (0.08)          (0.08)          (0.08)         
>  (0.08)          (0.09)          (0.10)          (0.12)          (0.16)   
L.lnroadden                -0.060          -0.137          -0.113          -0.104         
>  -0.138          -0.298**         0.166*         -0.057           0.570*  
                           (0.11)          (0.12)          (0.11)          (0.11)         
>  (0.12)          (0.15)          (0.10)          (0.12)          (0.34)   
reg_density                 0.061           0.302**         0.155**         0.054         
>   0.027           0.070           0.031           0.004           0.170** 
                           (0.08)          (0.13)          (0.08)          (0.08)         
>  (0.10)          (0.11)          (0.07)          (0.07)          (0.07)   
L.FDI inflows (% o~n        0.065           0.107           0.088           0.082         
>   0.066           0.072           0.181           0.332***        0.277   
                           (0.10)          (0.10)          (0.10)          (0.10)         
>  (0.10)          (0.11)          (0.12)          (0.12)          (0.19)   
(mean) rents                0.011           0.005           0.006           0.012         
>   0.012                          -0.000          -0.011           0.054** 
                           (0.01)          (0.01)          (0.01)          (0.01)         
>  (0.01)                          (0.01)          (0.01)          (0.02)   
(mean) pressfill            0.051           0.043           0.050           0.058         
>   0.070                           0.152           0.160          -0.105   
                           (0.09)          (0.09)          (0.09)          (0.09)         
>  (0.09)                          (0.10)          (0.10)          (0.17)   
(mean) KPRFmandate~e       -0.027**        -0.023**        -0.025**        -0.026**       
>  -0.026**                        -0.022*         -0.021*         -0.055***
                           (0.01)          (0.01)          (0.01)          (0.01)         
>  (0.01)                          (0.01)          (0.01)          (0.02)   
(mean) pctRussian           0.006           0.006           0.006           0.007         
>   0.003                           0.010**         0.005           0.023***
                           (0.01)          (0.01)          (0.01)          (0.01)         
>  (0.01)                          (0.01)          (0.01)          (0.01)   
L.Exports (% of GD~n                                                        0.180         
>                                                                           
                                                                           (0.20)         
>                                                                           
L.Exports (% of GD~a                                                       -0.023*        
>                                                                           
                                                                           (0.01)         
>                                                                           
L.Imports (% of GD~n                                                                      
>   0.018                                                                   
                                                                                          
>  (0.17)                                                                   
L.Imports (% of GD~a                                                                      
>   0.006                                                                   
                                                                                          
>  (0.01)                                                                   
Central                                                                                   
>                                   0.000                                   
                                                                                          
>                                     (.)                                   
Far East                                                                                  
>                                  -0.257                                   
                                                                                          
>                                  (0.44)                                   
North Caucasus                                                                            
>                                  -0.154                                   
                                                                                          
>                                  (0.48)                                   
North West                                                                                
>                                   0.935***                                
                                                                                          
>                                  (0.27)                                   
Siberia                                                                                   
>                                   0.428                                   
                                                                                          
>                                  (0.36)                                   
South                                                                                     
>                                   0.376                                   
                                                                                          
>                                  (0.33)                                   
Ural                                                                                      
>                                  -0.102                                   
                                                                                          
>                                  (0.38)                                   
Volga                                                                                     
>                                   0.483*                                  
                                                                                          
>                                  (0.25)                                   
Central                                                                                   
>                                                   0.000                   
                                                                                          
>                                                     (.)                   
Central Black Earth                                                                       
>                                                  -0.392                   
                                                                                          
>                                                  (0.36)                   
East Siberian                                                                             
>                                                  -0.757                   
                                                                                          
>                                                  (0.52)                   
Far Eastern                                                                               
>                                                  -1.101**                 
                                                                                          
>                                                  (0.54)                   
Kaliningrad                                                                               
>                                                   1.934***                
                                                                                          
>                                                  (0.55)                   
North Caucasus                                                                            
>                                                  -1.046**                 
                                                                                          
>                                                  (0.50)                   
Northern                                                                                  
>                                                   0.470                   
                                                                                          
>                                                  (0.38)                   
Northwestern                                                                              
>                                                   0.577                   
                                                                                          
>                                                  (0.37)                   
Ural                                                                                      
>                                                   0.235                   
                                                                                          
>                                                  (0.40)                   
Volga                                                                                     
>                                                   0.236                   
                                                                                          
>                                                  (0.33)                   
Volga-Vyatka                                                                              
>                                                   0.021                   
                                                                                          
>                                                  (0.33)                   
West Siberian                                                                             
>                                                   0.015                   
                                                                                          
>                                                  (0.46)                   
Constant                    5.313**         2.559           3.241           7.012***      
>   8.580***        9.308***       -2.371           0.060           1.888   
                           (2.28)          (2.27)          (2.31)          (2.14)         
>  (2.17)          (2.18)          (2.68)          (2.77)          (3.59)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
ln_r                                                                                      
>                                                                           
Constant                    0.377*          0.327*          0.387**         0.360*        
>   0.337*                                                          0.277   
                           (0.19)          (0.18)          (0.19)          (0.19)         
>  (0.19)                                                          (0.28)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
ln_s                                                                                      
>                                                                           
Constant                    1.208***        1.040***        1.168***        1.148***      
>   1.118***                                                        1.108** 
                           (0.27)          (0.25)          (0.26)          (0.27)         
>  (0.27)                                                          (0.43)   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
lnalpha                                                                                   
>                                                                           
Constant                                                                                  
>                                   0.032          -0.023                   
                                                                                          
>                                  (0.09)          (0.09)                   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------
# of observations             416             416             416             416         
>     416             419             416             416             209   
# of regions                   74              74              74              74         
>      74              72                                              37   
Prob > Chi2                 0.000           0.000           0.000           0.000         
>   0.000           0.000           0.000           0.000           0.000   
Log likelihood           -1221.65        -1213.72        -1213.22        -1220.44        -
> 1221.91         -876.75        -1262.37        -1251.72         -561.84   
AIC                       2483.31         2467.44         2466.44         2480.87         
> 2483.83         1779.49         2576.73         2563.43         1163.68   
------------------------------------------------------------------------------------------
> --------------------------------------------------------------------------

. 
. *** Figure 3 in Main Text
. xtnbreg protest_ikd2 $X2 $C1 $C2 if sample1==1, re

Fitting negative binomial (constant dispersion) model:

Iteration 0:   log likelihood = -74476.476  (not concave)
Iteration 1:   log likelihood = -62576.974  
Iteration 2:   log likelihood = -50630.648  (backed up)
Iteration 3:   log likelihood =  -36764.26  
Iteration 4:   log likelihood = -23441.835  
Iteration 5:   log likelihood = -6896.7262  
Iteration 6:   log likelihood =  -3446.534  
Iteration 7:   log likelihood = -2650.3626  
Iteration 8:   log likelihood = -2600.9612  
Iteration 9:   log likelihood =  -2600.016  
Iteration 10:  log likelihood = -2600.0137  
Iteration 11:  log likelihood = -2600.0137  

Iteration 0:   log likelihood = -3810.7631  
Iteration 1:   log likelihood = -1698.4354  
Iteration 2:   log likelihood = -1432.2595  
Iteration 3:   log likelihood =  -1431.161  
Iteration 4:   log likelihood = -1431.1601  
Iteration 5:   log likelihood = -1431.1601  

Iteration 0:   log likelihood = -1431.1601  (not concave)
Iteration 1:   log likelihood = -1390.6747  (not concave)
Iteration 2:   log likelihood = -1369.2858  
Iteration 3:   log likelihood = -1286.6513  
Iteration 4:   log likelihood = -1280.9839  
Iteration 5:   log likelihood =  -1280.922  
Iteration 6:   log likelihood =  -1280.922  

Fitting full model:

Iteration 0:   log likelihood = -1393.8367  (not concave)
Iteration 1:   log likelihood = -1376.1675  (not concave)
Iteration 2:   log likelihood =  -1266.408  
Iteration 3:   log likelihood = -1262.2387  
Iteration 4:   log likelihood = -1230.4718  
Iteration 5:   log likelihood = -1219.6794  
Iteration 6:   log likelihood = -1218.5542  
Iteration 7:   log likelihood = -1218.4897  
Iteration 8:   log likelihood = -1218.4895  
Iteration 9:   log likelihood = -1218.4895  

Random-effects negative binomial regression     Number of obs     =        416
Group variable: rcode                           Number of groups  =         74

Random effects u_i ~ Beta                       Obs per group:
                                                              min =          2
                                                              avg =        5.6
                                                              max =          6

                                                Wald chi2(17)     =      91.42
Log likelihood  = -1218.4895                    Prob > chi2       =     0.0000

----------------------------------------------------------------------------------------
          protest_ikd2 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
               lntrade |
                   L1. |   .5774318   .2440631     2.37   0.018     .0990768    1.055787
                       |
             educshare |   .0890607   .0571847     1.56   0.119    -.0230192    .2011406
                       |
cL.lntrade#c.educshare |  -.0524035   .0176025    -2.98   0.003    -.0869038   -.0179031
                       |
               reg_pop |
                   L1. |   .4833176   .0732753     6.60   0.000     .3397006    .6269346
                       |
        reg_urbanshare |
                   L1. |   .0136482   .0111266     1.23   0.220    -.0081596     .035456
                       |
               lngrppc |
                   L1. |  -.6144772   .1418066    -4.33   0.000    -.8924132   -.3365413
                       |
             reg_grpgr |
                   L1. |   -.428684   .2060136    -2.08   0.037    -.8324633   -.0249046
                       |
     reg_levelofunempl |
                   L1. |  -.0358944   .0295562    -1.21   0.225    -.0938236    .0220347
                       |
                lnnews |
                   L1. |  -.0669973   .1135012    -0.59   0.555    -.2894555     .155461
                       |
             lndistmos |
                   L1. |   .1791088   .0787578     2.27   0.023     .0247464    .3334712
                       |
             lnroadden |
                   L1. |  -.0786518   .1111613    -0.71   0.479     -.296524    .1392204
                       |
           reg_density |   .0703033   .0665109     1.06   0.291    -.0600555    .2006622
                       |
                 lnfdi |
                   L1. |     .07524   .0985714     0.76   0.445    -.1179563    .2684363
                       |
                 rents |   .0110966   .0095082     1.17   0.243    -.0075391    .0297322
             pressfill |   .0477562    .090624     0.53   0.598    -.1298636     .225376
      KPRFmandateshare |  -.0269951   .0109344    -2.47   0.014     -.048426   -.0055641
            pctRussian |   .0064228   .0061568     1.04   0.297    -.0056444      .01849
                 _cons |   4.593955   2.297061     2.00   0.046      .091799    9.096111
-----------------------+----------------------------------------------------------------
                 /ln_r |   .3869403     .19155                      .0115092    .7623714
                 /ln_s |   1.200743   .2685418                      .6744109    1.727075
-----------------------+----------------------------------------------------------------
                     r |   1.472469   .2820514                      1.011576    2.143353
                     s |   3.322585   .8922528                      1.962876     5.62418
----------------------------------------------------------------------------------------
LR test vs. pooled: chibar2(01) = 124.86               Prob >= chibar2 = 0.000

. est store m1

. margins, dydx(l.lntrade) at(educshare=(0(1)30))

Average marginal effects                        Number of obs     =        416
Model VCE    : OIM

Expression   : Linear prediction, predict()
dy/dx w.r.t. : L.lntrade

1._at        : educshare       =           0

2._at        : educshare       =           1

3._at        : educshare       =           2

4._at        : educshare       =           3

5._at        : educshare       =           4

6._at        : educshare       =           5

7._at        : educshare       =           6

8._at        : educshare       =           7

9._at        : educshare       =           8

10._at       : educshare       =           9

11._at       : educshare       =          10

12._at       : educshare       =          11

13._at       : educshare       =          12

14._at       : educshare       =          13

15._at       : educshare       =          14

16._at       : educshare       =          15

17._at       : educshare       =          16

18._at       : educshare       =          17

19._at       : educshare       =          18

20._at       : educshare       =          19

21._at       : educshare       =          20

22._at       : educshare       =          21

23._at       : educshare       =          22

24._at       : educshare       =          23

25._at       : educshare       =          24

26._at       : educshare       =          25

27._at       : educshare       =          26

28._at       : educshare       =          27

29._at       : educshare       =          28

30._at       : educshare       =          29

31._at       : educshare       =          30

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
L.lntrade    |
         _at |
          1  |   .5774318   .2440631     2.37   0.018     .0990768    1.055787
          2  |   .5250283   .2287008     2.30   0.022     .0767831    .9732735
          3  |   .4726248   .2136842     2.21   0.027     .0538114    .8914382
          4  |   .4202214   .1990918     2.11   0.035     .0300085    .8104342
          5  |   .3678179    .185024     1.99   0.047     .0051776    .7304582
          6  |   .3154144   .1716097     1.84   0.066    -.0209344    .6517632
          7  |   .2630109   .1590144     1.65   0.098    -.0486516    .5746735
          8  |   .2106075   .1474483     1.43   0.153    -.0783859    .4996008
          9  |    .158204   .1371718     1.15   0.249    -.1106478    .4270558
         10  |   .1058005   .1284948     0.82   0.410    -.1460447    .3576457
         11  |   .0533971   .1217597     0.44   0.661    -.1852475    .2920417
         12  |   .0009936   .1173014     0.01   0.993     -.228913    .2309002
         13  |  -.0514099   .1153843    -0.45   0.656    -.2775589    .1747391
         14  |  -.1038134   .1161341    -0.89   0.371     -.331432    .1238053
         15  |  -.1562168   .1195008    -1.31   0.191     -.390434    .0780004
         16  |  -.2086203   .1252734    -1.67   0.096    -.4541517    .0369111
         17  |  -.2610238   .1331395    -1.96   0.050    -.5219725   -.0000751
         18  |  -.3134272   .1427534    -2.20   0.028    -.5932188   -.0336357
         19  |  -.3658307   .1537876    -2.38   0.017    -.6672489   -.0644125
         20  |  -.4182342   .1659591    -2.52   0.012    -.7435081   -.0929603
         21  |  -.4706376   .1790361    -2.63   0.009     -.821542   -.1197333
         22  |  -.5230411   .1928344    -2.71   0.007    -.9009897   -.1450925
         23  |  -.5754446   .2072101    -2.78   0.005    -.9815689   -.1693203
         24  |  -.6278481   .2220509    -2.83   0.005     -1.06306   -.1926362
         25  |  -.6802515   .2372697    -2.87   0.004    -1.145292   -.2152115
         26  |   -.732655   .2527981    -2.90   0.004     -1.22813   -.2371798
         27  |  -.7850585   .2685826    -2.92   0.003    -1.311471   -.2586463
         28  |  -.8374619   .2845804    -2.94   0.003    -1.395229   -.2796947
         29  |  -.8898654   .3007575    -2.96   0.003    -1.479339   -.3003916
         30  |  -.9422689   .3170865    -2.97   0.003    -1.563747   -.3207907
         31  |  -.9946724   .3335451    -2.98   0.003    -1.648409   -.3409359
------------------------------------------------------------------------------

. marginsplot, level(90) ///
>         plotopts(msymbol(i) lcolor(black)) ///
>         ciopts(recast(rline) lpattern(dash) lcolor(black)) ///
>         title("") ///
>         xtitle("Share of population with secondary education (in %)", ///
>                 size(large) margin(medsmall)) ///
>         ytitle("Marginal effect of trade exposure", size(vlarge) margin(small)) ///
>         ylabel(, angle(hori) labsize(medsmall) nogrid) ///
>         xscale(noline range(-.5 26.5)) yscale(noline) ///
>         xlabel(1 "35" 6 "40" 11 "45" 16 "50" 21 "55" 26 "60", ///
>                 labsize(medsmall) nogrid) ///
>         yline(0, lcolor(black) lwidth(.2)) ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(ikd_trade, replace) xsize(3) ysize(2) ///
>         addplot(hist educshare, bin(21) yaxis(2) legend(off) ///
>                 ytitle("", axis(2)) ylabel(none, axis(2)) ///
>                 xlabel(1 "35" 6 "40" 11 "45" 16 "50" 21 "55" 26 "60", ///
>                         labsize(medsmall) nogrid) ///
>                 fcolor(none) lcolor(gs12))

  Variables that uniquely identify margins: educshare

. graph save "figure_effect", replace
(note: file figure_effect.gph not found)
(file figure_effect.gph saved)

. 
. 
. 
. ***** 5. Mechanism: regional level
. *------------------------------------------------------------------------------*
. 
. *** Model specification
. xtset rcode year
       panel variable:  rcode (strongly balanced)
        time variable:  year, 1990 to 2014
                delta:  1 unit

. global X cl.lntrade##c.educshare4 

. global C l.reg_pop l.reg_urbanshare l.lngrppc l.reg_grpgr ///
>         l.reg_levelofunempl l.lnnews l.lndistmos l.reg_density l.lnroadden

. 
. *** Average wage level (positive interaction)
. gen lnwage = ln(reg_avwage/reg_grp)
(763 missing values generated)

. xtpcse lnwage $X $C if year>=2000, pair

Number of gaps in sample:  13

Linear regression, correlated panels corrected standard errors (PCSEs)

Group variable:   rcode                         Number of obs     =        995
Time variable:    year                          Number of groups  =         80
Panels:           correlated (unbalanced)       Obs per group:
Autocorrelation:  no autocorrelation                          min =          3
Sigma computed by pairwise selection                          avg =    12.4375
                                                              max =         13
Estimated covariances      =      3240          R-squared         =     0.8217
Estimated autocorrelations =         0          Wald chi2(12)     =   48636.25
Estimated coefficients     =        13          Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
                        |           Panel-corrected
                 lnwage |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                lntrade |
                    L1. |  -.2828877   .0248331   -11.39   0.000    -.3315597   -.2342157
                        |
             educshare4 |  -.0363054   .0059194    -6.13   0.000    -.0479072   -.0247037
                        |
cL.lntrade#c.educshare4 |   .0164853   .0016111    10.23   0.000     .0133276    .0196431
                        |
                reg_pop |
                    L1. |  -.5551932    .005002  -110.99   0.000    -.5649969   -.5453894
                        |
         reg_urbanshare |
                    L1. |  -.0092061   .0009379    -9.82   0.000    -.0110445   -.0073678
                        |
                lngrppc |
                    L1. |   .0411161   .0288806     1.42   0.155    -.0154889    .0977211
                        |
              reg_grpgr |
                    L1. |   -.155685   .1104152    -1.41   0.159    -.3720948    .0607247
                        |
      reg_levelofunempl |
                    L1. |   .0206912   .0031553     6.56   0.000     .0145069    .0268754
                        |
                 lnnews |
                    L1. |  -.0392032   .0160633    -2.44   0.015    -.0706866   -.0077197
                        |
              lndistmos |
                    L1. |  -.0868318   .0061128   -14.21   0.000    -.0988126    -.074851
                        |
            reg_density |
                    L1. |   .1978259   .0162428    12.18   0.000     .1659907    .2296611
                        |
              lnroadden |
                    L1. |  -.0946624   .0150997    -6.27   0.000    -.1242573   -.0650676
                        |
                  _cons |   .3054002   .3503251     0.87   0.383    -.3812243    .9920247
-----------------------------------------------------------------------------------------

. est store mech1

. 
. *** Consumption per capita (positive interaction)
. gen cons = PERSCONSUMPC/1000
(885 missing values generated)

. xtpcse cons $X $C if year>=2000, pair

Number of gaps in sample:  8
(note: at least one disturbance covariance assumed 0, no common time periods
       between panels)

Linear regression, correlated panels corrected standard errors (PCSEs)

Group variable:   rcode                         Number of obs     =        690
Time variable:    year                          Number of groups  =         80
Panels:           correlated (unbalanced)       Obs per group:
Autocorrelation:  no autocorrelation                          min =          1
Sigma computed by pairwise selection                          avg =      8.625
                                                              max =          9
Estimated covariances      =      3240          R-squared         =     0.8393
Estimated autocorrelations =         0          Wald chi2(11)     =    7052.54
Estimated coefficients     =        13          Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
                        |           Panel-corrected
                   cons |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                lntrade |
                    L1. |   -.428738   .0782259    -5.48   0.000    -.5820581    -.275418
                        |
             educshare4 |  -.0676483   .0171809    -3.94   0.000    -.1013223   -.0339743
                        |
cL.lntrade#c.educshare4 |   .0256471   .0054302     4.72   0.000     .0150041    .0362902
                        |
                reg_pop |
                    L1. |   .0634926   .0312123     2.03   0.042     .0023177    .1246675
                        |
         reg_urbanshare |
                    L1. |  -.0055219   .0033174    -1.66   0.096    -.0120238    .0009801
                        |
                lngrppc |
                    L1. |    1.69915   .1578903    10.76   0.000     1.389691    2.008609
                        |
              reg_grpgr |
                    L1. |   .0816817   .4620288     0.18   0.860    -.8238781    .9872415
                        |
      reg_levelofunempl |
                    L1. |   .0097054   .0071466     1.36   0.174    -.0043017    .0237125
                        |
                 lnnews |
                    L1. |   .0396225   .0503463     0.79   0.431    -.0590544    .1382994
                        |
              lndistmos |
                    L1. |   .0683254   .0149364     4.57   0.000     .0390506    .0976002
                        |
            reg_density |
                    L1. |   .2391677   .0244099     9.80   0.000     .1913251    .2870103
                        |
              lnroadden |
                    L1. |   .1028454   .0652987     1.57   0.115    -.0251378    .2308286
                        |
                  _cons |  -16.37672   1.720143    -9.52   0.000    -19.74814    -13.0053
-----------------------------------------------------------------------------------------

. est store mech2

. 
. *** Level of employment (positive interaction)
. gen empl = (EMPL/(reg_pop*1000))
(1,018 missing values generated)

. xtpcse empl $X $C if year>=2000, pair

Number of gaps in sample:  9

Linear regression, correlated panels corrected standard errors (PCSEs)

Group variable:   rcode                         Number of obs     =        770
Time variable:    year                          Number of groups  =         80
Panels:           correlated (unbalanced)       Obs per group:
Autocorrelation:  no autocorrelation                          min =          2
Sigma computed by pairwise selection                          avg =      9.625
                                                              max =         10
Estimated covariances      =      3240          R-squared         =     0.5828
Estimated autocorrelations =         0          Wald chi2(10)     =    6449.40
Estimated coefficients     =        13          Prob > chi2       =     0.0000

-----------------------------------------------------------------------------------------
                        |           Panel-corrected
                   empl |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                lntrade |
                    L1. |  -.0145813   .0028105    -5.19   0.000    -.0200898   -.0090729
                        |
             educshare4 |  -.0038217   .0007989    -4.78   0.000    -.0053875   -.0022559
                        |
cL.lntrade#c.educshare4 |   .0011353   .0002472     4.59   0.000     .0006507    .0016199
                        |
                reg_pop |
                    L1. |  -.0110114   .0007145   -15.41   0.000    -.0124118   -.0096109
                        |
         reg_urbanshare |
                    L1. |   .0011947   .0000732    16.33   0.000     .0010514    .0013381
                        |
                lngrppc |
                    L1. |  -.0013295    .002826    -0.47   0.638    -.0068684    .0042093
                        |
              reg_grpgr |
                    L1. |  -.0128705   .0070384    -1.83   0.067    -.0266655    .0009245
                        |
      reg_levelofunempl |
                    L1. |  -.0056129   .0005597   -10.03   0.000    -.0067098    -.004516
                        |
                 lnnews |
                    L1. |   .0100807   .0017071     5.90   0.000     .0067347    .0134266
                        |
              lndistmos |
                    L1. |    .005897   .0007809     7.55   0.000     .0043664    .0074276
                        |
            reg_density |
                    L1. |   .0132487   .0005378    24.63   0.000     .0121946    .0143028
                        |
              lnroadden |
                    L1. |   -.002914   .0016284    -1.79   0.074    -.0061056    .0002775
                        |
                  _cons |   .4189743   .0376148    11.14   0.000     .3452507     .492698
-----------------------------------------------------------------------------------------

. est store mech3

. 
. *** Table 2 in Main Text
. esttab mech1 mech2 mech3, ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g r2 p, labels("# of observations" "# of regions" ///
>                 "R squared " "Prob > Chi2") fmt(0 0 2 3)) ///
>         order(*lntrade* educ*)

--------------------------------------------------------------------
                              (1)             (2)             (3)   
                           lnwage            cons            empl   
                             b/se            b/se            b/se   
--------------------------------------------------------------------
L.Imports+Exports ~n       -0.283***       -0.429***       -0.015***
                           (0.02)          (0.08)          (0.00)   
L.Imports+Exports ~#        0.016***        0.026***        0.001***
                           (0.00)          (0.01)          (0.00)   
educshare4                 -0.036***       -0.068***       -0.004***
                           (0.01)          (0.02)          (0.00)   
L.reg_pop                  -0.555***        0.063**        -0.011***
                           (0.01)          (0.03)          (0.00)   
L.reg_urbanshare           -0.009***       -0.006*          0.001***
                           (0.00)          (0.00)          (0.00)   
L.lngrppc                   0.041           1.699***       -0.001   
                           (0.03)          (0.16)          (0.00)   
L.reg_grpgr                -0.156           0.082          -0.013*  
                           (0.11)          (0.46)          (0.01)   
L.reg_levelofunempl         0.021***        0.010          -0.006***
                           (0.00)          (0.01)          (0.00)   
L.lnnews                   -0.039**         0.040           0.010***
                           (0.02)          (0.05)          (0.00)   
L.lndistmos                -0.087***        0.068***        0.006***
                           (0.01)          (0.01)          (0.00)   
L.reg_density               0.198***        0.239***        0.013***
                           (0.02)          (0.02)          (0.00)   
L.lnroadden                -0.095***        0.103          -0.003*  
                           (0.02)          (0.07)          (0.00)   
Constant                    0.305         -16.377***        0.419***
                           (0.35)          (1.72)          (0.04)   
--------------------------------------------------------------------
# of observations             995             690             770   
# of regions                   80              80              80   
R squared                    0.82            0.84            0.58   
Prob > Chi2                 0.000           0.000           0.000   
--------------------------------------------------------------------

. 
. 
. 
. ***** 6. Mechanism: individual risk
. *------------------------------------------------------------------------------*
. 
. *** Russian survey panel
. * Load data
. use "PRW_2020_ISQ_RLMSdata.dta", clear
( )

. * Restrict to 2000-2013
. drop if year<2000
(43,095 observations deleted)

. drop if year>2013
(36,802 observations deleted)

. * Keep adults only
. drop if child==1
(35,471 observations deleted)

. 
. *** Merge trade and FDI exposure
. * Generate merge variable
. gen conversion =        .               
(182,517 missing values generated)

. replace conversion =    1       if region ==    58
(5,136 real changes made)

. replace conversion =    1       if region ==    84
(3,564 real changes made)

. replace conversion =    2       if region ==    93
(4,891 real changes made)

. replace conversion =    3       if region ==    106
(3,853 real changes made)

. replace conversion =    3       if region ==    107
(4,419 real changes made)

. replace conversion =    4       if region ==    48
(4,979 real changes made)

. replace conversion =    5       if region ==    77
(6,641 real changes made)

. replace conversion =    6       if region ==    67
(3,941 real changes made)

. replace conversion =    7       if region ==    14
(3,417 real changes made)

. replace conversion =    8       if region ==    89
(4,065 real changes made)

. replace conversion =    8       if region ==    105
(3,916 real changes made)

. replace conversion =    9       if region ==    9
(3,730 real changes made)

. replace conversion =    9       if region ==    129
(6,312 real changes made)

. replace conversion =    10      if region ==    66
(4,351 real changes made)

. replace conversion =    10      if region ==    73
(4,241 real changes made)

. replace conversion =    11      if region ==    46
(3,482 real changes made)

. replace conversion =    12      if region ==    1
(3,903 real changes made)

. replace conversion =    13      if region ==    72
(5,716 real changes made)

. replace conversion =    14      if region ==    138
(15,337 real changes made)

. replace conversion =    14      if region ==    140
(0 real changes made)

. replace conversion =    15      if region ==    142
(8,506 real changes made)

. replace conversion =    16      if region ==    116
(4,184 real changes made)

. replace conversion =    17      if region ==    161
(2,867 real changes made)

. replace conversion =    18      if region ==    47
(4,501 real changes made)

. replace conversion =    19      if region ==    117
(4,278 real changes made)

. replace conversion =    20      if region ==    12
(4,476 real changes made)

. replace conversion =    21      if region ==    92
(3,532 real changes made)

. replace conversion =    22      if region ==    137
(4,477 real changes made)

. replace conversion =    23      if region ==    141
(5,949 real changes made)

. replace conversion =    24      if region ==    70
(4,609 real changes made)

. replace conversion =    24      if region ==    100
(3,730 real changes made)

. replace conversion =    25      if region ==    135
(3,788 real changes made)

. replace conversion =    26      if region ==    52
(4,200 real changes made)

. replace conversion =    27      if region ==    33
(3,917 real changes made)

. replace conversion =    28      if region ==    45
(4,913 real changes made)

. replace conversion =    29      if region ==    71
(4,287 real changes made)

. replace conversion =    30      if region ==    136
(4,046 real changes made)

. replace conversion =    31      if region ==    86
(397 real changes made)

. replace conversion =    32      if region ==    10
(4,843 real changes made)

. replace conversion =    33      if region ==    39
(5,123 real changes made)

. label define conversion ///                                     
>         1 "Altai K"     2 "Amur" 3 "Chelyabinsk" 4 "Chuvash" 5 "Kabardino-Balkar" ///
>         6 "Kaliningrad" 7 "Kaluga" 8 "Komi" 9 "Krasnodar" 10 "Krasnoyarsk" ///
>         11 "Kurgan"     12 "Leningrad" 13 "Lipetsk"     14 "Moscow C" 15 "Moscow O" ///
>         16 "Nizhny Novgorod" 17 "Novosibirsk" 18 "Orenburg" 19 "Penza" 20 "Perm" ///
>         21 "Primorsky" 22 "Rostov" 23 "Saint Petersburg" 24 "Saratov" ///
>         25 "Smolensk" 26 "Stavropol" 27 "Tambov" 28 "Tatarstan" 29 "Tomsk" ///
>         30 "Tula" 31 "Tyumen" 32 "Udmurt" 33 "Volgograd"                

. label values conversion conversion                                      

. 
. *** Merge trade and FDI
. merge m:m conversion year using "glob_stats.dta"
(note: variable year was int, now double to accommodate using data's values)

    Result                           # of obs.
    -----------------------------------------
    not matched                           714
        from master                         0  (_merge==1)
        from using                        714  (_merge==2)

    matched                           182,517  (_merge==3)
    -----------------------------------------

. drop if _merge==2
(714 observations deleted)

. drop _merge

. 
. *** Education level
. gen educlvl = diplom-1 if diplom<=6
(414 missing values generated)

. 
. *** Dependent variables
. * Satisfaction with economic condition
. gen satecon = j66_1

. replace satecon = . if j66_1>5
(1,434 real changes made, 1,434 to missing)

. * Concern about getting necessities
. gen necess = j66

. replace necess = . if j66>5
(2,082 real changes made, 2,082 to missing)

. * Rank on social ladder
. gen econlad = j62

. replace econlad = . if j62>9
(3,686 real changes made, 3,686 to missing)

. gen powlad = j63

. replace powlad = . if j63>9
(5,599 real changes made, 5,599 to missing)

. gen ladder = (econlad + powlad)/2
(6,393 missing values generated)

. gen ladder_rev = (ladder*(-1))+10
(6,393 missing values generated)

. * Unemployed and would like to find work
. gen unemp = 0 if j1<=5
(130 missing values generated)

. replace unemp = 1 if j1==5
(82,932 real changes made)

. gen unempwork = 0 if unemp!=.
(130 missing values generated)

. replace unempwork = 1 if unemp==1 & j81==1
(20,744 real changes made)

. 
. *** Control variables
. * Gender
. gen gender = 0 if h5==1
(104,819 missing values generated)

. replace gender = 1 if h5==2
(104,819 real changes made)

. * Age
. replace age = . if age>104
(18 real changes made, 18 to missing)

. * Income
. gen income = j60

. replace income = . if j60>2265000
(5,249 real changes made, 5,249 to missing)

. gen lnincome = ln(income+1)
(5,249 missing values generated)

. * Marriage
. gen married = 0 if j72_17<=6 | j322<=6
(426 missing values generated)

. replace married = 1 if j72_17==2 | j322==2 | j322==3
(90,813 real changes made)

. * Second job
. gen secjob = j32 if j32<=2
(83,204 missing values generated)

. replace secjob = 0 if secjob==2
(94,987 real changes made)

. replace secjob = 0 if unemp==1
(82,932 real changes made)

. * Works for government
. gen govfirm = j23 if j23<=2
(95,309 missing values generated)

. replace govfirm = 0 if govfirm==2
(39,897 real changes made)

. replace govfirm = 0 if unemp==1
(82,932 real changes made)

. * Works for foreign firm
. gen forfirm = j24 if j24<=2
(95,326 missing values generated)

. replace forfirm = 0 if forfirm==2
(83,881 real changes made)

. replace forfirm = 0 if unemp==1
(82,932 real changes made)

. * Works for Russian firm
. gen rusfirm = j25 if j25<=2
(96,610 missing values generated)

. replace rusfirm = 0 if rusfirm==2
(41,271 real changes made)

. replace rusfirm = 0 if unemp==1
(82,932 real changes made)

. * Self-employed
. gen selfemp = j26 if j26<=2
(92,353 missing values generated)

. replace selfemp = 0 if selfemp==2
(86,329 real changes made)

. replace selfemp = 0 if unemp==1
(82,932 real changes made)

. * Pension
. gen pension = j73 if j73<=2
(93 missing values generated)

. replace pension = 0 if pension==2
(124,131 real changes made)

. 
. *** Model specification
. global X1 c.educlvl##c.lntrade i.year

. global X2 c.educlvl##c.lnfdi i.year

. global C1 gender age lnincome married i.urban secjob selfemp govfirm

. xtset idind year
       panel variable:  idind (unbalanced)
        time variable:  year, 2000 to 2013, but with gaps
                delta:  1 unit

. 
. *** Regression models
. * Economic dissatisfaction
. meoprobit satecon $X1 $C1 if pension!=1 || idind:

Fitting fixed-effects model:

Iteration 0:   log likelihood = -153235.58  
Iteration 1:   log likelihood = -150317.79  
Iteration 2:   log likelihood = -150317.02  
Iteration 3:   log likelihood = -150317.02  

Refining starting values:

Grid node 0:   log likelihood =  -146354.4

Fitting full model:

Iteration 0:   log likelihood =  -146354.4  (not concave)
Iteration 1:   log likelihood = -144971.21  
Iteration 2:   log likelihood = -142430.45  
Iteration 3:   log likelihood = -142271.72  
Iteration 4:   log likelihood = -142268.24  
Iteration 5:   log likelihood = -142268.23  

Mixed-effects oprobit regression                Number of obs     =    107,242
Group variable:           idind                 Number of groups  =     26,076

                                                Obs per group:
                                                              min =          1
                                                              avg =        4.1
                                                              max =         14

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(26)     =    3279.27
Log likelihood = -142268.23                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
            satecon |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .0744322   .0182587     4.08   0.000     .0386459    .1102185
            lntrade |   .0765841   .0187899     4.08   0.000     .0397565    .1134116
                    |
c.educlvl#c.lntrade |  -.0115786   .0051544    -2.25   0.025    -.0216811   -.0014761
                    |
               year |
              2001  |  -.1575868   .0217302    -7.25   0.000    -.2001772   -.1149963
              2002  |  -.2663975   .0215854   -12.34   0.000    -.3087042   -.2240908
              2003  |  -.2744083   .0217359   -12.62   0.000    -.3170099   -.2318067
              2004  |   -.284281   .0218124   -13.03   0.000    -.3270325   -.2415295
              2005  |  -.3705205   .0219459   -16.88   0.000    -.4135337   -.3275073
              2006  |  -.3305064   .0215901   -15.31   0.000    -.3728222   -.2881905
              2007  |  -.3596057   .0218704   -16.44   0.000    -.4024708   -.3167405
              2008  |  -.4482842   .0220774   -20.31   0.000    -.4915551   -.4050132
              2009  |  -.3302401   .0224635   -14.70   0.000    -.3742679   -.2862124
              2010  |  -.4065826   .0212711   -19.11   0.000    -.4482732   -.3648919
              2011  |  -.4249502   .0214154   -19.84   0.000    -.4669236   -.3829767
              2012  |  -.5132144   .0213005   -24.09   0.000    -.5549626   -.4714662
              2013  |   -.449874   .0217291   -20.70   0.000    -.4924622   -.4072858
                    |
             gender |   .0264729   .0129438     2.05   0.041     .0011034    .0518424
                age |   .0214248   .0005584    38.37   0.000     .0203303    .0225193
           lnincome |  -.0378425   .0013122   -28.84   0.000    -.0404144   -.0352706
            married |  -.1292545   .0113322   -11.41   0.000    -.1514652   -.1070438
                    |
              urban |
              City  |   .1053153   .0163066     6.46   0.000     .0733549    .1372757
        Small town  |   .0995892   .0284739     3.50   0.000     .0437814    .1553971
           Village  |   .2575805   .0174477    14.76   0.000     .2233836    .2917775
                    |
             secjob |   -.092919   .0235687    -3.94   0.000    -.1391128   -.0467252
            selfemp |  -.2346897   .0262547    -8.94   0.000     -.286148   -.1832313
            govfirm |  -.0018242   .0103702    -0.18   0.860    -.0221495     .018501
--------------------+----------------------------------------------------------------
              /cut1 |  -1.933792   .0720488   -26.84   0.000    -2.075005   -1.792579
              /cut2 |  -.5821991   .0715907    -8.13   0.000    -.7225143   -.4418839
              /cut3 |   .1279676   .0715576     1.79   0.074    -.0122826    .2682179
              /cut4 |    1.31847   .0716185    18.41   0.000       1.1781    1.458839
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .6140133   .0108339                      .5931422    .6356189
-------------------------------------------------------------------------------------
LR test vs. oprobit model: chibar2(01) = 16097.57     Prob >= chibar2 = 0.0000

. est store satecon1

. meoprobit satecon $X1 lnfdi $C1 if pension!=1 || idind:
note: 2000.year identifies no observations in the sample
note: 2001.year identifies no observations in the sample
note: 2002.year identifies no observations in the sample
note: 2003.year identifies no observations in the sample
note: 2013.year omitted because of collinearity

Fitting fixed-effects model:

Iteration 0:   log likelihood = -114594.19  
Iteration 1:   log likelihood = -112464.48  
Iteration 2:   log likelihood = -112463.91  
Iteration 3:   log likelihood = -112463.91  

Refining starting values:

Grid node 0:   log likelihood =  -109666.4

Fitting full model:

Iteration 0:   log likelihood =  -109666.4  (not concave)
Iteration 1:   log likelihood = -108639.51  
Iteration 2:   log likelihood = -106346.87  
Iteration 3:   log likelihood = -106276.82  
Iteration 4:   log likelihood = -106275.85  
Iteration 5:   log likelihood = -106275.85  

Mixed-effects oprobit regression                Number of obs     =     79,849
Group variable:           idind                 Number of groups  =     22,488

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.6
                                                              max =         10

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(23)     =    2461.74
Log likelihood = -106275.85                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
            satecon |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .0859533    .023639     3.64   0.000     .0396217    .1322849
            lntrade |   .0496087    .024377     2.04   0.042     .0018307    .0973868
                    |
c.educlvl#c.lntrade |  -.0201533     .00663    -3.04   0.002    -.0331478   -.0071588
                    |
               year |
              2000  |          0  (empty)
              2001  |          0  (empty)
              2002  |          0  (empty)
              2003  |          0  (empty)
              2004  |   .1546742   .0202139     7.65   0.000     .1150557    .1942927
              2005  |   .0610641   .0201286     3.03   0.002     .0216127    .1005154
              2006  |   .0429761   .0189136     2.27   0.023     .0059063     .080046
              2007  |   .0787935    .018956     4.16   0.000     .0416405    .1159465
              2008  |  -.0309408   .0189629    -1.63   0.103    -.0681074    .0062258
              2009  |   .1042718   .0189968     5.49   0.000     .0670387    .1415048
              2010  |   .0390065   .0167437     2.33   0.020     .0061893    .0718236
              2011  |   .0307809    .016697     1.84   0.065    -.0019446    .0635065
              2012  |  -.0701676   .0162633    -4.31   0.000     -.102043   -.0382921
              2013  |          0  (omitted)
                    |
              lnfdi |  -.0070949   .0106278    -0.67   0.504     -.027925    .0137352
             gender |   .0227188   .0146214     1.55   0.120    -.0059386    .0513761
                age |   .0235025   .0006487    36.23   0.000     .0222311    .0247739
           lnincome |  -.0402131   .0015574   -25.82   0.000    -.0432656   -.0371606
            married |  -.1545529   .0132731   -11.64   0.000    -.1805677   -.1285381
                    |
              urban |
              City  |   .0805839   .0183755     4.39   0.000     .0445685    .1165993
        Small town  |   .1575592   .0338797     4.65   0.000     .0911562    .2239622
           Village  |   .3167987   .0196965    16.08   0.000     .2781941    .3554032
                    |
             secjob |  -.1415297   .0277444    -5.10   0.000    -.1959078   -.0871516
            selfemp |  -.3819744   .0362139   -10.55   0.000    -.4529523   -.3109966
            govfirm |  -.0441768   .0124291    -3.55   0.000    -.0685374   -.0198161
--------------------+----------------------------------------------------------------
              /cut1 |  -1.740437   .0902671   -19.28   0.000    -1.917357   -1.563517
              /cut2 |  -.3383227   .0898505    -3.77   0.000    -.5144264    -.162219
              /cut3 |    .404412   .0898747     4.50   0.000     .2282608    .5805633
              /cut4 |   1.615944   .0900362    17.95   0.000     1.439476    1.792411
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .6753439   .0130866                      .6501755    .7014865
-------------------------------------------------------------------------------------
LR test vs. oprobit model: chibar2(01) = 12376.13     Prob >= chibar2 = 0.0000

. est store satecon2

. * Concerned about getting necessities
. gen necess_rev = (necess*(-1))+6
(2,082 missing values generated)

. meoprobit necess_rev $X1 $C1 if pension!=1 || idind:

Fitting fixed-effects model:

Iteration 0:   log likelihood = -153496.54  
Iteration 1:   log likelihood = -149977.05  
Iteration 2:   log likelihood = -149976.04  
Iteration 3:   log likelihood = -149976.04  

Refining starting values:

Grid node 0:   log likelihood = -144974.24

Fitting full model:

Iteration 0:   log likelihood = -144974.24  
Iteration 1:   log likelihood = -143910.98  
Iteration 2:   log likelihood = -141500.68  
Iteration 3:   log likelihood = -141474.72  
Iteration 4:   log likelihood = -141474.62  
Iteration 5:   log likelihood = -141474.62  

Mixed-effects oprobit regression                Number of obs     =    106,893
Group variable:           idind                 Number of groups  =     26,031

                                                Obs per group:
                                                              min =          1
                                                              avg =        4.1
                                                              max =         14

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(26)     =    2987.36
Log likelihood = -141474.62                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
         necess_rev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .1536323   .0188066     8.17   0.000     .1167719    .1904926
            lntrade |   .0870947   .0193867     4.49   0.000     .0490974     .125092
                    |
c.educlvl#c.lntrade |  -.0353293   .0053104    -6.65   0.000    -.0457375    -.024921
                    |
               year |
              2001  |  -.1194423   .0224098    -5.33   0.000    -.1633647   -.0755198
              2002  |  -.1205495   .0222666    -5.41   0.000    -.1641911   -.0769078
              2003  |  -.1534815   .0224585    -6.83   0.000    -.1974993   -.1094636
              2004  |  -.1998314   .0224742    -8.89   0.000      -.24388   -.1557827
              2005  |  -.2047481   .0226716    -9.03   0.000    -.2491835   -.1603127
              2006  |   -.286462   .0222432   -12.88   0.000    -.3300578   -.2428661
              2007  |  -.3189225   .0225185   -14.16   0.000    -.3630579   -.2747872
              2008  |  -.3073285   .0227779   -13.49   0.000    -.3519724   -.2626846
              2009  |  -.3631974   .0230703   -15.74   0.000    -.4084144   -.3179805
              2010  |  -.3832467   .0219107   -17.49   0.000    -.4261909   -.3403024
              2011  |  -.4152076   .0220264   -18.85   0.000    -.4583785   -.3720366
              2012  |   -.405774   .0219455   -18.49   0.000    -.4487864   -.3627616
              2013  |  -.3936127   .0223535   -17.61   0.000    -.4374248   -.3498005
                    |
             gender |   .1925553   .0134382    14.33   0.000     .1662169    .2188936
                age |   .0189505   .0005783    32.77   0.000     .0178172    .0200839
           lnincome |  -.0099785   .0013367    -7.46   0.000    -.0125985   -.0073586
            married |   .0474828   .0117157     4.05   0.000     .0245205    .0704451
                    |
              urban |
              City  |   .2847369   .0169489    16.80   0.000     .2515175    .3179562
        Small town  |   .2096116   .0295804     7.09   0.000     .1516351    .2675882
           Village  |    .471999   .0181943    25.94   0.000     .4363388    .5076592
                    |
             secjob |  -.1042897   .0242772    -4.30   0.000    -.1518721   -.0567073
            selfemp |   -.197965   .0272653    -7.26   0.000     -.251404   -.1445261
            govfirm |  -.0222916   .0107168    -2.08   0.038    -.0432961    -.001287
--------------------+----------------------------------------------------------------
              /cut1 |  -.9792974   .0739881   -13.24   0.000    -1.124311   -.8342834
              /cut2 |  -.0767811   .0738266    -1.04   0.298    -.2214785    .0679163
              /cut3 |   .4102202   .0738181     5.56   0.000     .2655394    .5549011
              /cut4 |    1.49973   .0739028    20.29   0.000     1.354883    1.644577
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .6691696   .0118423                      .6463571    .6927872
-------------------------------------------------------------------------------------
LR test vs. oprobit model: chibar2(01) = 17002.84     Prob >= chibar2 = 0.0000

. est store necess1

. meoprobit necess_rev $X1 lnfdi $C1 if pension!=1 || idind:
note: 2000.year identifies no observations in the sample
note: 2001.year identifies no observations in the sample
note: 2002.year identifies no observations in the sample
note: 2003.year identifies no observations in the sample
note: 2013.year omitted because of collinearity

Fitting fixed-effects model:

Iteration 0:   log likelihood = -115343.97  
Iteration 1:   log likelihood = -112764.24  
Iteration 2:   log likelihood =  -112763.6  
Iteration 3:   log likelihood =  -112763.6  

Refining starting values:

Grid node 0:   log likelihood = -108713.17

Fitting full model:

Iteration 0:   log likelihood = -108713.17  
Iteration 1:   log likelihood = -107987.21  
Iteration 2:   log likelihood = -105930.75  
Iteration 3:   log likelihood =  -105665.9  
Iteration 4:   log likelihood = -105658.09  
Iteration 5:   log likelihood = -105658.09  

Mixed-effects oprobit regression                Number of obs     =     79,657
Group variable:           idind                 Number of groups  =     22,447

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.5
                                                              max =         10

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(23)     =    2231.50
Log likelihood = -105658.09                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
         necess_rev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .1947068   .0247704     7.86   0.000     .1461578    .2432558
            lntrade |   .1297887   .0255849     5.07   0.000     .0796433    .1799341
                    |
c.educlvl#c.lntrade |  -.0496435   .0069484    -7.14   0.000     -.063262   -.0360249
                    |
               year |
              2000  |          0  (empty)
              2001  |          0  (empty)
              2002  |          0  (empty)
              2003  |          0  (empty)
              2004  |    .157066   .0208785     7.52   0.000     .1161449     .197987
              2005  |   .1670186   .0208623     8.01   0.000     .1261293    .2079079
              2006  |   .0590086   .0195011     3.03   0.002     .0207872    .0972301
              2007  |   .0994626   .0195499     5.09   0.000     .0611456    .1377796
              2008  |   .0790565   .0195828     4.04   0.000     .0406749    .1174381
              2009  |   .0645778   .0194746     3.32   0.001     .0264084    .1027473
              2010  |   .0208364   .0172156     1.21   0.226    -.0129056    .0545785
              2011  |   -.024279    .017098    -1.42   0.156    -.0577904    .0092324
              2012  |  -.0139209   .0166966    -0.83   0.404    -.0466456    .0188037
              2013  |          0  (omitted)
                    |
              lnfdi |  -.0429747    .011057    -3.89   0.000    -.0646461   -.0213033
             gender |   .1894995   .0156005    12.15   0.000      .158923     .220076
                age |   .0211224   .0006882    30.69   0.000     .0197735    .0224712
           lnincome |  -.0091274   .0015928    -5.73   0.000    -.0122492   -.0060055
            married |   .0307818   .0139637     2.20   0.027     .0034135    .0581501
                    |
              urban |
              City  |   .2715886   .0196253    13.84   0.000     .2331237    .3100535
        Small town  |   .2040419   .0361786     5.64   0.000     .1331331    .2749508
           Village  |   .5127683   .0210934    24.31   0.000      .471426    .5541107
                    |
             secjob |  -.1187342   .0288211    -4.12   0.000    -.1752226   -.0622458
            selfemp |  -.2296406   .0378515    -6.07   0.000    -.3038282    -.155453
            govfirm |  -.0428479   .0129804    -3.30   0.001     -.068289   -.0174068
--------------------+----------------------------------------------------------------
              /cut1 |  -.5500677   .0944543    -5.82   0.000    -.7351948   -.3649407
              /cut2 |   .4029697   .0943372     4.27   0.000     .2180722    .5878672
              /cut3 |   .9126294   .0943636     9.67   0.000     .7276801    1.097579
              /cut4 |   2.072446   .0945521    21.92   0.000     1.887127    2.257765
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .8000562   .0152578                      .7707033    .8305269
-------------------------------------------------------------------------------------
LR test vs. oprobit model: chibar2(01) = 14211.03     Prob >= chibar2 = 0.0000

. est store necess2

. * Unemployed and wants to find work
. meprobit unempwork $X1 $C1 if pension!=1 || idind:
note: secjob != 0 predicts failure perfectly
      secjob dropped and 3300 obs not used

note: selfemp != 0 predicts failure perfectly
      selfemp dropped and 2839 obs not used

note: govfirm != 0 predicts failure perfectly
      govfirm dropped and 34677 obs not used


Fitting fixed-effects model:

Iteration 0:   log likelihood = -34225.406  
Iteration 1:   log likelihood = -34181.295  
Iteration 2:   log likelihood = -34181.276  
Iteration 3:   log likelihood = -34181.276  

Refining starting values:

Grid node 0:   log likelihood = -31896.327

Fitting full model:

Iteration 0:   log likelihood = -31896.327  
Iteration 1:   log likelihood = -31643.268  
Iteration 2:   log likelihood = -31435.411  
Iteration 3:   log likelihood = -31426.796  
Iteration 4:   log likelihood = -31426.772  
Iteration 5:   log likelihood = -31426.772  

Mixed-effects probit regression                 Number of obs     =     67,274
Group variable:           idind                 Number of groups  =     20,220

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.3
                                                              max =         14

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(23)     =    4877.47
Log likelihood = -31426.772                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
          unempwork |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .4398587   .0305598    14.39   0.000     .3799627    .4997547
            lntrade |   .1684828   .0296933     5.67   0.000     .1102849    .2266807
                    |
c.educlvl#c.lntrade |   -.077722   .0086945    -8.94   0.000     -.094763    -.060681
                    |
               year |
              2001  |  -.0172958   .0402221    -0.43   0.667    -.0961297    .0615381
              2002  |  -.1319659   .0399592    -3.30   0.001    -.2102845   -.0536473
              2003  |  -.2082717   .0400197    -5.20   0.000    -.2867089   -.1298346
              2004  |  -.2465509   .0398577    -6.19   0.000    -.3246706   -.1684312
              2005  |  -.4029526   .0402508   -10.01   0.000    -.4818427   -.3240624
              2006  |  -.4365959   .0397441   -10.99   0.000    -.5144929   -.3586989
              2007  |  -.5816992   .0408813   -14.23   0.000    -.6618252   -.5015732
              2008  |   -.658921   .0415456   -15.86   0.000    -.7403488   -.5774932
              2009  |  -.5564272    .041431   -13.43   0.000    -.6376304    -.475224
              2010  |  -.6137454   .0390943   -15.70   0.000    -.6903689    -.537122
              2011  |  -.5886186   .0392669   -14.99   0.000    -.6655803    -.511657
              2012  |  -.6411892   .0392269   -16.35   0.000    -.7180726   -.5643059
              2013  |  -.7933666   .0404222   -19.63   0.000    -.8725928   -.7141404
                    |
             gender |  -.0129283   .0210467    -0.61   0.539     -.054179    .0283225
                age |   .0103758   .0009252    11.21   0.000     .0085625    .0121892
           lnincome |   -.113613   .0020373   -55.77   0.000    -.1176059     -.10962
            married |  -.2092365   .0208105   -10.05   0.000    -.2500244   -.1684486
                    |
              urban |
              City  |  -.1323815   .0273154    -4.85   0.000    -.1859187   -.0788443
        Small town  |   .0480115   .0457523     1.05   0.294    -.0416613    .1376843
           Village  |   .2387236   .0282561     8.45   0.000     .1833426    .2941045
                    |
             secjob |          0  (omitted)
            selfemp |          0  (omitted)
            govfirm |          0  (omitted)
              _cons |  -1.234978   .1143772   -10.80   0.000    -1.459153   -1.010802
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .9102394   .0284821                      .8560929    .9678106
-------------------------------------------------------------------------------------
LR test vs. probit model: chibar2(01) = 5509.01       Prob >= chibar2 = 0.0000

. est store unempwork1

. meprobit unempwork $X1 lnfdi $C1 if pension!=1 || idind:
note: secjob != 0 predicts failure perfectly
      secjob dropped and 2496 obs not used

note: selfemp != 0 predicts failure perfectly
      selfemp dropped and 1541 obs not used

note: govfirm != 0 predicts failure perfectly
      govfirm dropped and 25195 obs not used

note: 2000.year identifies no observations in the sample
note: 2001.year identifies no observations in the sample
note: 2002.year identifies no observations in the sample
note: 2003.year identifies no observations in the sample
note: 2013.year omitted because of collinearity

Fitting fixed-effects model:

Iteration 0:   log likelihood = -24159.288  
Iteration 1:   log likelihood = -24098.143  
Iteration 2:   log likelihood = -24098.081  
Iteration 3:   log likelihood = -24098.081  

Refining starting values:

Grid node 0:   log likelihood = -22556.498

Fitting full model:

Iteration 0:   log likelihood = -22556.498  
Iteration 1:   log likelihood = -22247.644  
Iteration 2:   log likelihood = -22116.615  
Iteration 3:   log likelihood = -22113.487  
Iteration 4:   log likelihood = -22113.484  
Iteration 5:   log likelihood = -22113.484  

Mixed-effects probit regression                 Number of obs     =     51,291
Group variable:           idind                 Number of groups  =     17,359

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.0
                                                              max =         10

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(20)     =    3245.70
Log likelihood = -22113.484                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
          unempwork |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .3882485   .0420593     9.23   0.000     .3058139    .4706832
            lntrade |   .1643118   .0405365     4.05   0.000     .0848617    .2437618
                    |
c.educlvl#c.lntrade |  -.0703426   .0118254    -5.95   0.000    -.0935199   -.0471653
                    |
               year |
              2000  |          0  (empty)
              2001  |          0  (empty)
              2002  |          0  (empty)
              2003  |          0  (empty)
              2004  |   .5389421    .039203    13.75   0.000     .4621057    .6157785
              2005  |   .3602104   .0394591     9.13   0.000      .282872    .4375488
              2006  |     .31644   .0378673     8.36   0.000     .2422214    .3906586
              2007  |   .2254685   .0385703     5.85   0.000      .149872    .3010649
              2008  |   .1405974   .0392464     3.58   0.000      .063676    .2175189
              2009  |   .2847359   .0383818     7.42   0.000     .2095089    .3599628
              2010  |   .1941525     .03457     5.62   0.000     .1263965    .2619085
              2011  |   .2068703   .0345251     5.99   0.000     .1392024    .2745382
              2012  |   .1455068   .0339719     4.28   0.000     .0789231    .2120905
              2013  |          0  (omitted)
                    |
              lnfdi |  -.0321773   .0201568    -1.60   0.110     -.071684    .0073293
             gender |  -.0017181   .0247672    -0.07   0.945    -.0502608    .0468247
                age |   .0105812   .0011017     9.60   0.000      .008422    .0127405
           lnincome |  -.1252607   .0024591   -50.94   0.000    -.1300805   -.1204409
            married |  -.2550721    .025221   -10.11   0.000    -.3045042   -.2056399
                    |
              urban |
              City  |  -.1888719   .0320542    -5.89   0.000    -.2516969   -.1260469
        Small town  |  -.2117961    .059878    -3.54   0.000    -.3291549   -.0944373
           Village  |   .2537775   .0329044     7.71   0.000     .1892861    .3182689
                    |
             secjob |          0  (omitted)
            selfemp |          0  (omitted)
            govfirm |          0  (omitted)
              _cons |  -1.860993   .1501056   -12.40   0.000    -2.155195   -1.566792
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   1.038391   .0381979                        .96616    1.116023
-------------------------------------------------------------------------------------
LR test vs. probit model: chibar2(01) = 3969.19       Prob >= chibar2 = 0.0000

. est store unempwork2

. * Rank on social ladder
. mixed ladder_rev $X1 $C1 if pension!=1 || idind:

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log likelihood = -168658.89  
Iteration 1:   log likelihood = -168658.88  

Computing standard errors:

Mixed-effects ML regression                     Number of obs     =    104,695
Group variable: idind                           Number of groups  =     25,857

                                                Obs per group:
                                                              min =          1
                                                              avg =        4.0
                                                              max =         14

                                                Wald chi2(26)     =    3366.28
Log likelihood = -168658.88                     Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
         ladder_rev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |  -.0275683    .018585    -1.48   0.138    -.0639942    .0088577
            lntrade |   .1566388   .0191895     8.16   0.000      .119028    .1942496
                    |
c.educlvl#c.lntrade |  -.0174728   .0052447    -3.33   0.001    -.0277522   -.0071935
                    |
               year |
              2001  |  -.2104814    .020908   -10.07   0.000    -.2514602   -.1695025
              2002  |  -.1315396   .0207434    -6.34   0.000     -.172196   -.0908832
              2003  |  -.1596546   .0209667    -7.61   0.000    -.2007486   -.1185606
              2004  |  -.3516265   .0210266   -16.72   0.000    -.3928379   -.3104151
              2005  |  -.3013152   .0212248   -14.20   0.000     -.342915   -.2597154
              2006  |  -.3213136   .0209392   -15.35   0.000    -.3623537   -.2802734
              2007  |  -.3567755   .0211638   -16.86   0.000    -.3982558   -.3152953
              2008  |  -.4349116   .0214413   -20.28   0.000    -.4769357   -.3928875
              2009  |  -.3127307   .0218132   -14.34   0.000    -.3554839   -.2699776
              2010  |  -.3742229   .0206857   -18.09   0.000    -.4147661   -.3336797
              2011  |  -.3658957   .0208526   -17.55   0.000     -.406766   -.3250253
              2012  |  -.5096101   .0207547   -24.55   0.000    -.5502885   -.4689317
              2013  |  -.4755211   .0211669   -22.47   0.000    -.5170074   -.4340347
                    |
             gender |   .0765922   .0134094     5.71   0.000     .0503103    .1028742
                age |   .0224556   .0005682    39.52   0.000      .021342    .0235693
           lnincome |  -.0241682   .0012797   -18.89   0.000    -.0266764   -.0216601
            married |  -.1212751   .0113531   -10.68   0.000    -.1435267   -.0990234
                    |
              urban |
              City  |   .0921834   .0168705     5.46   0.000      .059118    .1252489
        Small town  |  -.0116518   .0295906    -0.39   0.694    -.0696482    .0463446
           Village  |   .0671636   .0180304     3.73   0.000     .0318247    .1025025
                    |
             secjob |  -.0233802   .0232788    -1.00   0.315    -.0690059    .0222455
            selfemp |  -.2248488   .0260342    -8.64   0.000    -.2758749   -.1738227
            govfirm |  -.0804989   .0102826    -7.83   0.000    -.1006524   -.0603454
              _cons |   5.561834   .0729487    76.24   0.000     5.418857    5.704811
-------------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
idind: Identity              |
                  var(_cons) |   .7046538   .0100408      .6852464    .7246109
-----------------------------+------------------------------------------------
               var(Residual) |   1.116433   .0056017      1.105507    1.127466
------------------------------------------------------------------------------
LR test vs. linear model: chibar2(01) = 20192.60      Prob >= chibar2 = 0.0000

. est store ladder1

. mixed ladder_rev $X1 lnfdi $C1 if pension!=1 || idind:

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log likelihood = -124495.87  
Iteration 1:   log likelihood = -124495.87  

Computing standard errors:

Mixed-effects ML regression                     Number of obs     =     78,035
Group variable: idind                           Number of groups  =     22,311

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.5
                                                              max =         10

                                                Wald chi2(23)     =    2167.62
Log likelihood = -124495.87                     Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
         ladder_rev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |  -.0580155   .0232827    -2.49   0.013    -.1036487   -.0123823
            lntrade |   .0349682   .0240637     1.45   0.146    -.0121957    .0821321
                    |
c.educlvl#c.lntrade |  -.0109438   .0065303    -1.68   0.094     -.023743    .0018553
                    |
               year |
              2005  |   .0482121   .0190841     2.53   0.012      .010808    .0856163
              2006  |  -.0311908   .0187711    -1.66   0.097    -.0679814    .0055998
              2007  |  -.0313037   .0192167    -1.63   0.103    -.0689677    .0063603
              2008  |  -.1123808   .0196514    -5.72   0.000    -.1508969   -.0738647
              2009  |  -.0347048   .0198879    -1.75   0.081    -.0736843    .0042748
              2010  |  -.0835481   .0185406    -4.51   0.000     -.119887   -.0472091
              2011  |  -.0554393   .0187281    -2.96   0.003    -.0921457    -.018733
              2012  |  -.1858803   .0186695    -9.96   0.000    -.2224719   -.1492887
              2013  |  -.1565673    .019087    -8.20   0.000    -.1939772   -.1191574
                    |
              lnfdi |  -.0054153   .0101818    -0.53   0.595    -.0253713    .0145407
             gender |   .0780108   .0146529     5.32   0.000     .0492916      .10673
                age |   .0218782   .0006386    34.26   0.000     .0206266    .0231298
           lnincome |  -.0233437   .0014609   -15.98   0.000     -.026207   -.0204805
            married |    -.13686    .012898   -10.61   0.000    -.1621395   -.1115805
                    |
              urban |
              City  |   .0500614   .0183969     2.72   0.007     .0140041    .0861188
        Small town  |   .1437074   .0340698     4.22   0.000     .0769318    .2104829
           Village  |   .1197918   .0196737     6.09   0.000     .0812321    .1583516
                    |
             secjob |  -.0474525   .0264958    -1.79   0.073    -.0993832    .0044783
            selfemp |   -.392194   .0349475   -11.22   0.000    -.4606899   -.3236981
            govfirm |  -.1170777   .0119326    -9.81   0.000     -.140465   -.0936903
              _cons |   5.758896   .0885642    65.03   0.000     5.585314    5.932479
-------------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
idind: Identity              |
                  var(_cons) |   .7313188   .0111631      .7097636    .7535287
-----------------------------+------------------------------------------------
               var(Residual) |   1.035393   .0061833      1.023344    1.047583
------------------------------------------------------------------------------
LR test vs. linear model: chibar2(01) = 15344.52      Prob >= chibar2 = 0.0000

. est store ladder2

. 
. *** Table A2 in the Online Appendix
. tabstat unempwork necess_rev satecon ladder_rev lntrade educlvl lnfdi ///
>         gender age lnincome married secjob selfemp govfirm urban ///
>         if pension!=1, statistics(N mean sd min max) format(%9.2f)

   stats |  unempw~k  necess~v   satecon  ladder~v   lntrade   educlvl     lnfdi    gender
---------+--------------------------------------------------------------------------------
       N | 124153.00 122752.00 123215.00 120086.00 123932.00 124007.00  92396.00 124224.00
    mean |      0.14      3.77      3.66      6.00      3.39      3.40      0.67      0.53
      sd |      0.35      1.25      1.13      1.39      0.84      1.15      0.56      0.50
     min |      0.00      1.00      1.00      1.00      0.92      0.00      0.00      0.00
     max |      1.00      5.00      5.00      9.00      5.13      5.00      3.13      1.00
------------------------------------------------------------------------------------------

   stats |       age  lnincome   married    secjob   selfemp   govfirm     urban
---------+----------------------------------------------------------------------
       N | 124220.00 119571.00 123906.00 124026.00 115585.00 113010.00 124224.00
    mean |     33.61      6.93      0.51      0.03      0.03      0.34      2.18
      sd |     11.81      3.71      0.50      0.17      0.17      0.48      1.22
     min |     13.00      0.00      0.00      0.00      0.00      0.00      1.00
     max |    101.00     14.63      1.00      1.00      1.00      1.00      4.00
--------------------------------------------------------------------------------

. 
. *** Table 3 in Main Text
. esttab unempwork2 necess2 satecon2 ladder2, ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g p ll aic, labels("# of observations" "# of regions" ///
>                 "Prob > Chi2" "Log likelihood" "AIC") fmt(0 0 3 2 2)) ///
>         order(*lntrade* educ*)

------------------------------------------------------------------------------------
                              (1)             (2)             (3)             (4)   
                        unempwork      necess_rev         satecon      ladder_rev   
                             b/se            b/se            b/se            b/se   
------------------------------------------------------------------------------------
main                                                                                
Imports+Exports (%~n        0.164***        0.130***        0.050**         0.035   
                           (0.04)          (0.03)          (0.02)          (0.02)   
educlvl # Imports+~        -0.070***       -0.050***       -0.020***       -0.011*  
                           (0.01)          (0.01)          (0.01)          (0.01)   
educlvl                     0.388***        0.195***        0.086***       -0.058** 
                           (0.04)          (0.02)          (0.02)          (0.02)   
Survey year=2000            0.000           0.000           0.000                   
                              (.)             (.)             (.)                   
Survey year=2001            0.000           0.000           0.000                   
                              (.)             (.)             (.)                   
Survey year=2002            0.000           0.000           0.000                   
                              (.)             (.)             (.)                   
Survey year=2003            0.000           0.000           0.000                   
                              (.)             (.)             (.)                   
Survey year=2004            0.539***        0.157***        0.155***        0.000   
                           (0.04)          (0.02)          (0.02)             (.)   
Survey year=2005            0.360***        0.167***        0.061***        0.048** 
                           (0.04)          (0.02)          (0.02)          (0.02)   
Survey year=2006            0.316***        0.059***        0.043**        -0.031*  
                           (0.04)          (0.02)          (0.02)          (0.02)   
Survey year=2007            0.225***        0.099***        0.079***       -0.031   
                           (0.04)          (0.02)          (0.02)          (0.02)   
Survey year=2008            0.141***        0.079***       -0.031          -0.112***
                           (0.04)          (0.02)          (0.02)          (0.02)   
Survey year=2009            0.285***        0.065***        0.104***       -0.035*  
                           (0.04)          (0.02)          (0.02)          (0.02)   
Survey year=2010            0.194***        0.021           0.039**        -0.084***
                           (0.03)          (0.02)          (0.02)          (0.02)   
Survey year=2011            0.207***       -0.024           0.031*         -0.055***
                           (0.03)          (0.02)          (0.02)          (0.02)   
Survey year=2012            0.146***       -0.014          -0.070***       -0.186***
                           (0.03)          (0.02)          (0.02)          (0.02)   
Survey year=2013            0.000           0.000           0.000          -0.157***
                              (.)             (.)             (.)          (0.02)   
FDI inflows (% of ~n       -0.032          -0.043***       -0.007          -0.005   
                           (0.02)          (0.01)          (0.01)          (0.01)   
gender                     -0.002           0.189***        0.023           0.078***
                           (0.02)          (0.02)          (0.01)          (0.01)   
NUMBER OF FULL YEARS        0.011***        0.021***        0.024***        0.022***
                           (0.00)          (0.00)          (0.00)          (0.00)   
lnincome                   -0.125***       -0.009***       -0.040***       -0.023***
                           (0.00)          (0.00)          (0.00)          (0.00)   
married                    -0.255***        0.031**        -0.155***       -0.137***
                           (0.03)          (0.01)          (0.01)          (0.01)   
Regional center             0.000           0.000           0.000           0.000   
                              (.)             (.)             (.)             (.)   
City                       -0.189***        0.272***        0.081***        0.050***
                           (0.03)          (0.02)          (0.02)          (0.02)   
Small town                 -0.212***        0.204***        0.158***        0.144***
                           (0.06)          (0.04)          (0.03)          (0.03)   
Village                     0.254***        0.513***        0.317***        0.120***
                           (0.03)          (0.02)          (0.02)          (0.02)   
secjob                      0.000          -0.119***       -0.142***       -0.047*  
                              (.)          (0.03)          (0.03)          (0.03)   
selfemp                     0.000          -0.230***       -0.382***       -0.392***
                              (.)          (0.04)          (0.04)          (0.03)   
govfirm                     0.000          -0.043***       -0.044***       -0.117***
                              (.)          (0.01)          (0.01)          (0.01)   
Constant                   -1.861***                                        5.759***
                           (0.15)                                          (0.09)   
------------------------------------------------------------------------------------
var(_cons[idind])                                                                   
Constant                    1.038***        0.800***        0.675***                
                           (0.04)          (0.02)          (0.01)                   
------------------------------------------------------------------------------------
cut1                                                                                
Constant                                   -0.550***       -1.740***                
                                           (0.09)          (0.09)                   
------------------------------------------------------------------------------------
cut2                                                                                
Constant                                    0.403***       -0.338***                
                                           (0.09)          (0.09)                   
------------------------------------------------------------------------------------
cut3                                                                                
Constant                                    0.913***        0.404***                
                                           (0.09)          (0.09)                   
------------------------------------------------------------------------------------
cut4                                                                                
Constant                                    2.072***        1.616***                
                                           (0.09)          (0.09)                   
------------------------------------------------------------------------------------
lns1_1_1                                                                            
Constant                                                                   -0.156***
                                                                           (0.01)   
------------------------------------------------------------------------------------
lnsig_e                                                                             
Constant                                                                    0.017***
                                                                           (0.00)   
------------------------------------------------------------------------------------
# of observations           51291           79657           79849           78035   
# of regions                                                                        
Prob > Chi2                 0.000           0.000           0.000           0.000   
Log likelihood          -22113.48      -105658.09      -106275.85      -124495.87   
AIC                      44270.97       211372.17       212607.70       249043.74   
------------------------------------------------------------------------------------

. 
. *** Figure
. meprobit unempwork $X1 lnfdi $C1 if pension!=1 || idind:
note: secjob != 0 predicts failure perfectly
      secjob dropped and 2496 obs not used

note: selfemp != 0 predicts failure perfectly
      selfemp dropped and 1541 obs not used

note: govfirm != 0 predicts failure perfectly
      govfirm dropped and 25195 obs not used

note: 2000.year identifies no observations in the sample
note: 2001.year identifies no observations in the sample
note: 2002.year identifies no observations in the sample
note: 2003.year identifies no observations in the sample
note: 2013.year omitted because of collinearity

Fitting fixed-effects model:

Iteration 0:   log likelihood = -24159.288  
Iteration 1:   log likelihood = -24098.143  
Iteration 2:   log likelihood = -24098.081  
Iteration 3:   log likelihood = -24098.081  

Refining starting values:

Grid node 0:   log likelihood = -22556.498

Fitting full model:

Iteration 0:   log likelihood = -22556.498  
Iteration 1:   log likelihood = -22247.644  
Iteration 2:   log likelihood = -22116.615  
Iteration 3:   log likelihood = -22113.487  
Iteration 4:   log likelihood = -22113.484  
Iteration 5:   log likelihood = -22113.484  

Mixed-effects probit regression                 Number of obs     =     51,291
Group variable:           idind                 Number of groups  =     17,359

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.0
                                                              max =         10

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(20)     =    3245.70
Log likelihood = -22113.484                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
          unempwork |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .3882485   .0420593     9.23   0.000     .3058139    .4706832
            lntrade |   .1643118   .0405365     4.05   0.000     .0848617    .2437618
                    |
c.educlvl#c.lntrade |  -.0703426   .0118254    -5.95   0.000    -.0935199   -.0471653
                    |
               year |
              2000  |          0  (empty)
              2001  |          0  (empty)
              2002  |          0  (empty)
              2003  |          0  (empty)
              2004  |   .5389421    .039203    13.75   0.000     .4621057    .6157785
              2005  |   .3602104   .0394591     9.13   0.000      .282872    .4375488
              2006  |     .31644   .0378673     8.36   0.000     .2422214    .3906586
              2007  |   .2254685   .0385703     5.85   0.000      .149872    .3010649
              2008  |   .1405974   .0392464     3.58   0.000      .063676    .2175189
              2009  |   .2847359   .0383818     7.42   0.000     .2095089    .3599628
              2010  |   .1941525     .03457     5.62   0.000     .1263965    .2619085
              2011  |   .2068703   .0345251     5.99   0.000     .1392024    .2745382
              2012  |   .1455068   .0339719     4.28   0.000     .0789231    .2120905
              2013  |          0  (omitted)
                    |
              lnfdi |  -.0321773   .0201568    -1.60   0.110     -.071684    .0073293
             gender |  -.0017181   .0247672    -0.07   0.945    -.0502608    .0468247
                age |   .0105812   .0011017     9.60   0.000      .008422    .0127405
           lnincome |  -.1252607   .0024591   -50.94   0.000    -.1300805   -.1204409
            married |  -.2550721    .025221   -10.11   0.000    -.3045042   -.2056399
                    |
              urban |
              City  |  -.1888719   .0320542    -5.89   0.000    -.2516969   -.1260469
        Small town  |  -.2117961    .059878    -3.54   0.000    -.3291549   -.0944373
           Village  |   .2537775   .0329044     7.71   0.000     .1892861    .3182689
                    |
             secjob |          0  (omitted)
            selfemp |          0  (omitted)
            govfirm |          0  (omitted)
              _cons |  -1.860993   .1501056   -12.40   0.000    -2.155195   -1.566792
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   1.038391   .0381979                        .96616    1.116023
-------------------------------------------------------------------------------------
LR test vs. probit model: chibar2(01) = 3969.19       Prob >= chibar2 = 0.0000

. margins, dydx(lntrade) at(educlvl=(0(1)5))

Average marginal effects                        Number of obs     =     51,291
Model VCE    : OIM

Expression   : Marginal predicted mean, predict()
dy/dx w.r.t. : lntrade

1._at        : educlvl         =           0

2._at        : educlvl         =           1

3._at        : educlvl         =           2

4._at        : educlvl         =           3

5._at        : educlvl         =           4

6._at        : educlvl         =           5

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
lntrade      |
         _at |
          1  |    .023358   .0057916     4.03   0.000     .0120066    .0347093
          2  |   .0147238   .0047434     3.10   0.002     .0054269    .0240206
          3  |   .0040459   .0036932     1.10   0.273    -.0031927    .0112844
          4  |  -.0086633   .0031872    -2.72   0.007    -.0149101   -.0024164
          5  |  -.0232819   .0040036    -5.82   0.000    -.0311287    -.015435
          6  |  -.0395746   .0059368    -6.67   0.000    -.0512105   -.0279387
------------------------------------------------------------------------------

. marginsplot, level(95) ///
>         plotopts(msymbol(i) lcolor(black)) ///
>         ciopts(recast(rline) lpattern(dash) lcolor(black)) ///
>         title("") ///
>         xtitle("Individual education level", size(large) margin(medsmall)) ///
>         ytitle("Marginal effect of trade exposure", size(vlarge) margin(small)) ///
>         ylabel(, angle(hori) labsize(medsmall) nogrid) ///
>         xscale(noline range(-.5 5.5)) yscale(noline) ///
>         xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                 2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                 4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "', ///
>                         labsize(medsmall) nogrid) ///
>         yline(0, lcolor(black) lwidth(.2)) ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(unemp_trade, replace) ///
>         addplot(hist educlvl, discrete yaxis(2) legend(off) ///
>                 ytitle("", axis(2)) ylabel(none, axis(2)) ///
>                 xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                         2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                         4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "',
>  ///
>                                 labsize(medsmall) nogrid) ///
>                 fcolor(none) lcolor(gs12))

  Variables that uniquely identify margins: educlvl

. 
. *** Figure
. meoprobit necess_rev $X1 lnfdi $C1 if pension!=1 || idind:
note: 2000.year identifies no observations in the sample
note: 2001.year identifies no observations in the sample
note: 2002.year identifies no observations in the sample
note: 2003.year identifies no observations in the sample
note: 2013.year omitted because of collinearity

Fitting fixed-effects model:

Iteration 0:   log likelihood = -115343.97  
Iteration 1:   log likelihood = -112764.24  
Iteration 2:   log likelihood =  -112763.6  
Iteration 3:   log likelihood =  -112763.6  

Refining starting values:

Grid node 0:   log likelihood = -108713.17

Fitting full model:

Iteration 0:   log likelihood = -108713.17  
Iteration 1:   log likelihood = -107987.21  
Iteration 2:   log likelihood = -105930.75  
Iteration 3:   log likelihood =  -105665.9  
Iteration 4:   log likelihood = -105658.09  
Iteration 5:   log likelihood = -105658.09  

Mixed-effects oprobit regression                Number of obs     =     79,657
Group variable:           idind                 Number of groups  =     22,447

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.5
                                                              max =         10

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(23)     =    2231.50
Log likelihood = -105658.09                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
         necess_rev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .1947068   .0247704     7.86   0.000     .1461578    .2432558
            lntrade |   .1297887   .0255849     5.07   0.000     .0796433    .1799341
                    |
c.educlvl#c.lntrade |  -.0496435   .0069484    -7.14   0.000     -.063262   -.0360249
                    |
               year |
              2000  |          0  (empty)
              2001  |          0  (empty)
              2002  |          0  (empty)
              2003  |          0  (empty)
              2004  |    .157066   .0208785     7.52   0.000     .1161449     .197987
              2005  |   .1670186   .0208623     8.01   0.000     .1261293    .2079079
              2006  |   .0590086   .0195011     3.03   0.002     .0207872    .0972301
              2007  |   .0994626   .0195499     5.09   0.000     .0611456    .1377796
              2008  |   .0790565   .0195828     4.04   0.000     .0406749    .1174381
              2009  |   .0645778   .0194746     3.32   0.001     .0264084    .1027473
              2010  |   .0208364   .0172156     1.21   0.226    -.0129056    .0545785
              2011  |   -.024279    .017098    -1.42   0.156    -.0577904    .0092324
              2012  |  -.0139209   .0166966    -0.83   0.404    -.0466456    .0188037
              2013  |          0  (omitted)
                    |
              lnfdi |  -.0429747    .011057    -3.89   0.000    -.0646461   -.0213033
             gender |   .1894995   .0156005    12.15   0.000      .158923     .220076
                age |   .0211224   .0006882    30.69   0.000     .0197735    .0224712
           lnincome |  -.0091274   .0015928    -5.73   0.000    -.0122492   -.0060055
            married |   .0307818   .0139637     2.20   0.027     .0034135    .0581501
                    |
              urban |
              City  |   .2715886   .0196253    13.84   0.000     .2331237    .3100535
        Small town  |   .2040419   .0361786     5.64   0.000     .1331331    .2749508
           Village  |   .5127683   .0210934    24.31   0.000      .471426    .5541107
                    |
             secjob |  -.1187342   .0288211    -4.12   0.000    -.1752226   -.0622458
            selfemp |  -.2296406   .0378515    -6.07   0.000    -.3038282    -.155453
            govfirm |  -.0428479   .0129804    -3.30   0.001     -.068289   -.0174068
--------------------+----------------------------------------------------------------
              /cut1 |  -.5500677   .0944543    -5.82   0.000    -.7351948   -.3649407
              /cut2 |   .4029697   .0943372     4.27   0.000     .2180722    .5878672
              /cut3 |   .9126294   .0943636     9.67   0.000     .7276801    1.097579
              /cut4 |   2.072446   .0945521    21.92   0.000     1.887127    2.257765
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .8000562   .0152578                      .7707033    .8305269
-------------------------------------------------------------------------------------
LR test vs. oprobit model: chibar2(01) = 14211.03     Prob >= chibar2 = 0.0000

. margins, dydx(lntrade) at(educlvl=(0(1)5)) predict(outcome(5))

Average marginal effects                        Number of obs     =     79,657
Model VCE    : OIM

Expression   : Marginal predicted mean (5.necess_rev), predict(outcome(5))
dy/dx w.r.t. : lntrade

1._at        : educlvl         =           0

2._at        : educlvl         =           1

3._at        : educlvl         =           2

4._at        : educlvl         =           3

5._at        : educlvl         =           4

6._at        : educlvl         =           5

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
lntrade      |
         _at |
          1  |   .0333819   .0065728     5.08   0.000     .0204994    .0462644
          2  |    .020795   .0050471     4.12   0.000     .0109029    .0306871
          3  |   .0079762   .0036857     2.16   0.030     .0007523       .0152
          4  |  -.0050397   .0028272    -1.78   0.075    -.0105808    .0005015
          5  |  -.0182145   .0030024    -6.07   0.000    -.0240992   -.0123299
          6  |  -.0315079   .0040979    -7.69   0.000    -.0395395   -.0234762
------------------------------------------------------------------------------

. marginsplot, level(95) ///
>         plotopts(msymbol(i) lcolor(black)) ///
>         ciopts(recast(rline) lpattern(dash) lcolor(black)) ///
>         title("") ///
>         xtitle("Individual education level", size(large) margin(medsmall)) ///
>         ytitle("Marginal effect of trade exposure", size(vlarge) margin(small)) ///
>         ylabel(, angle(hori) labsize(medsmall) nogrid) ///
>         xscale(noline range(-.5 5.5)) yscale(noline) ///
>         xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                 2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                 4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "', ///
>                         labsize(medsmall) nogrid) ///
>         yline(0, lcolor(black) lwidth(.2)) ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(necess_trade, replace) ///
>         addplot(hist educlvl, discrete yaxis(2) legend(off) ///
>                 ytitle("", axis(2)) ylabel(none, axis(2)) ///
>                 xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                         2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                         4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "',
>  ///
>                                 labsize(medsmall) nogrid) ///
>                 fcolor(none) lcolor(gs12))

  Variables that uniquely identify margins: educlvl

.         
. *** Figure
. meoprobit satecon $X1 lnfdi $C1 if pension!=1 || idind:
note: 2000.year identifies no observations in the sample
note: 2001.year identifies no observations in the sample
note: 2002.year identifies no observations in the sample
note: 2003.year identifies no observations in the sample
note: 2013.year omitted because of collinearity

Fitting fixed-effects model:

Iteration 0:   log likelihood = -114594.19  
Iteration 1:   log likelihood = -112464.48  
Iteration 2:   log likelihood = -112463.91  
Iteration 3:   log likelihood = -112463.91  

Refining starting values:

Grid node 0:   log likelihood =  -109666.4

Fitting full model:

Iteration 0:   log likelihood =  -109666.4  (not concave)
Iteration 1:   log likelihood = -108639.51  
Iteration 2:   log likelihood = -106346.87  
Iteration 3:   log likelihood = -106276.82  
Iteration 4:   log likelihood = -106275.85  
Iteration 5:   log likelihood = -106275.85  

Mixed-effects oprobit regression                Number of obs     =     79,849
Group variable:           idind                 Number of groups  =     22,488

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.6
                                                              max =         10

Integration method: mvaghermite                 Integration pts.  =          7

                                                Wald chi2(23)     =    2461.74
Log likelihood = -106275.85                     Prob > chi2       =     0.0000
-------------------------------------------------------------------------------------
            satecon |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |   .0859533    .023639     3.64   0.000     .0396217    .1322849
            lntrade |   .0496087    .024377     2.04   0.042     .0018307    .0973868
                    |
c.educlvl#c.lntrade |  -.0201533     .00663    -3.04   0.002    -.0331478   -.0071588
                    |
               year |
              2000  |          0  (empty)
              2001  |          0  (empty)
              2002  |          0  (empty)
              2003  |          0  (empty)
              2004  |   .1546742   .0202139     7.65   0.000     .1150557    .1942927
              2005  |   .0610641   .0201286     3.03   0.002     .0216127    .1005154
              2006  |   .0429761   .0189136     2.27   0.023     .0059063     .080046
              2007  |   .0787935    .018956     4.16   0.000     .0416405    .1159465
              2008  |  -.0309408   .0189629    -1.63   0.103    -.0681074    .0062258
              2009  |   .1042718   .0189968     5.49   0.000     .0670387    .1415048
              2010  |   .0390065   .0167437     2.33   0.020     .0061893    .0718236
              2011  |   .0307809    .016697     1.84   0.065    -.0019446    .0635065
              2012  |  -.0701676   .0162633    -4.31   0.000     -.102043   -.0382921
              2013  |          0  (omitted)
                    |
              lnfdi |  -.0070949   .0106278    -0.67   0.504     -.027925    .0137352
             gender |   .0227188   .0146214     1.55   0.120    -.0059386    .0513761
                age |   .0235025   .0006487    36.23   0.000     .0222311    .0247739
           lnincome |  -.0402131   .0015574   -25.82   0.000    -.0432656   -.0371606
            married |  -.1545529   .0132731   -11.64   0.000    -.1805677   -.1285381
                    |
              urban |
              City  |   .0805839   .0183755     4.39   0.000     .0445685    .1165993
        Small town  |   .1575592   .0338797     4.65   0.000     .0911562    .2239622
           Village  |   .3167987   .0196965    16.08   0.000     .2781941    .3554032
                    |
             secjob |  -.1415297   .0277444    -5.10   0.000    -.1959078   -.0871516
            selfemp |  -.3819744   .0362139   -10.55   0.000    -.4529523   -.3109966
            govfirm |  -.0441768   .0124291    -3.55   0.000    -.0685374   -.0198161
--------------------+----------------------------------------------------------------
              /cut1 |  -1.740437   .0902671   -19.28   0.000    -1.917357   -1.563517
              /cut2 |  -.3383227   .0898505    -3.77   0.000    -.5144264    -.162219
              /cut3 |    .404412   .0898747     4.50   0.000     .2282608    .5805633
              /cut4 |   1.615944   .0900362    17.95   0.000     1.439476    1.792411
--------------------+----------------------------------------------------------------
idind               |
          var(_cons)|   .6753439   .0130866                      .6501755    .7014865
-------------------------------------------------------------------------------------
LR test vs. oprobit model: chibar2(01) = 12376.13     Prob >= chibar2 = 0.0000

. margins, dydx(lntrade) at(educlvl=(0(1)5)) predict(outcome(5))

Average marginal effects                        Number of obs     =     79,849
Model VCE    : OIM

Expression   : Marginal predicted mean (5.satecon), predict(outcome(5))
dy/dx w.r.t. : lntrade

1._at        : educlvl         =           0

2._at        : educlvl         =           1

3._at        : educlvl         =           2

4._at        : educlvl         =           3

5._at        : educlvl         =           4

6._at        : educlvl         =           5

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
lntrade      |
         _at |
          1  |   .0117826   .0057885     2.04   0.042     .0004372    .0231279
          2  |   .0070565    .004433     1.59   0.111    -.0016319     .015745
          3  |   .0022472   .0032279     0.70   0.486    -.0040793    .0085737
          4  |  -.0026428   .0024662    -1.07   0.284    -.0074764    .0021908
          5  |  -.0076107    .002627    -2.90   0.004    -.0127595    -.002462
          6  |  -.0126535   .0036223    -3.49   0.000    -.0197531   -.0055539
------------------------------------------------------------------------------

. marginsplot, level(95) ///
>         plotopts(msymbol(i) lcolor(black)) ///
>         ciopts(recast(rline) lpattern(dash) lcolor(black)) ///
>         title("") ///
>         xtitle("Individual education level", size(large) margin(medsmall)) ///
>         ytitle("Marginal effect of trade exposure", size(vlarge) margin(small)) ///
>         ylabel(, angle(hori) labsize(medsmall) nogrid) ///
>         xscale(noline range(-.5 5.5)) yscale(noline) ///
>         xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                 2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                 4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "', ///
>                         labsize(medsmall) nogrid) ///
>         yline(0, lcolor(black) lwidth(.2)) ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(satecon_trade, replace) ///
>         addplot(hist educlvl, discrete yaxis(2) legend(off) ///
>                 ytitle("", axis(2)) ylabel(none, axis(2)) ///
>                 xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                         2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                         4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "',
>  ///
>                                 labsize(medsmall) nogrid) ///
>                 fcolor(none) lcolor(gs12))

  Variables that uniquely identify margins: educlvl

. 
. *** Figure
. mixed ladder_rev $X1 lnfdi $C1 if pension!=1 || idind:

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log likelihood = -124495.87  
Iteration 1:   log likelihood = -124495.87  

Computing standard errors:

Mixed-effects ML regression                     Number of obs     =     78,035
Group variable: idind                           Number of groups  =     22,311

                                                Obs per group:
                                                              min =          1
                                                              avg =        3.5
                                                              max =         10

                                                Wald chi2(23)     =    2167.62
Log likelihood = -124495.87                     Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
         ladder_rev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
            educlvl |  -.0580155   .0232827    -2.49   0.013    -.1036487   -.0123823
            lntrade |   .0349682   .0240637     1.45   0.146    -.0121957    .0821321
                    |
c.educlvl#c.lntrade |  -.0109438   .0065303    -1.68   0.094     -.023743    .0018553
                    |
               year |
              2005  |   .0482121   .0190841     2.53   0.012      .010808    .0856163
              2006  |  -.0311908   .0187711    -1.66   0.097    -.0679814    .0055998
              2007  |  -.0313037   .0192167    -1.63   0.103    -.0689677    .0063603
              2008  |  -.1123808   .0196514    -5.72   0.000    -.1508969   -.0738647
              2009  |  -.0347048   .0198879    -1.75   0.081    -.0736843    .0042748
              2010  |  -.0835481   .0185406    -4.51   0.000     -.119887   -.0472091
              2011  |  -.0554393   .0187281    -2.96   0.003    -.0921457    -.018733
              2012  |  -.1858803   .0186695    -9.96   0.000    -.2224719   -.1492887
              2013  |  -.1565673    .019087    -8.20   0.000    -.1939772   -.1191574
                    |
              lnfdi |  -.0054153   .0101818    -0.53   0.595    -.0253713    .0145407
             gender |   .0780108   .0146529     5.32   0.000     .0492916      .10673
                age |   .0218782   .0006386    34.26   0.000     .0206266    .0231298
           lnincome |  -.0233437   .0014609   -15.98   0.000     -.026207   -.0204805
            married |    -.13686    .012898   -10.61   0.000    -.1621395   -.1115805
                    |
              urban |
              City  |   .0500614   .0183969     2.72   0.007     .0140041    .0861188
        Small town  |   .1437074   .0340698     4.22   0.000     .0769318    .2104829
           Village  |   .1197918   .0196737     6.09   0.000     .0812321    .1583516
                    |
             secjob |  -.0474525   .0264958    -1.79   0.073    -.0993832    .0044783
            selfemp |   -.392194   .0349475   -11.22   0.000    -.4606899   -.3236981
            govfirm |  -.1170777   .0119326    -9.81   0.000     -.140465   -.0936903
              _cons |   5.758896   .0885642    65.03   0.000     5.585314    5.932479
-------------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
idind: Identity              |
                  var(_cons) |   .7313188   .0111631      .7097636    .7535287
-----------------------------+------------------------------------------------
               var(Residual) |   1.035393   .0061833      1.023344    1.047583
------------------------------------------------------------------------------
LR test vs. linear model: chibar2(01) = 15344.52      Prob >= chibar2 = 0.0000

. margins, dydx(lntrade) at(educlvl=(0(1)5))

Average marginal effects                        Number of obs     =     78,035

Expression   : Linear prediction, fixed portion, predict()
dy/dx w.r.t. : lntrade

1._at        : educlvl         =           0

2._at        : educlvl         =           1

3._at        : educlvl         =           2

4._at        : educlvl         =           3

5._at        : educlvl         =           4

6._at        : educlvl         =           5

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
lntrade      |
         _at |
          1  |   .0349682   .0240637     1.45   0.146    -.0121957    .0821321
          2  |   .0240244   .0182823     1.31   0.189    -.0118082    .0598569
          3  |   .0130805   .0132178     0.99   0.322    -.0128259     .038987
          4  |   .0021367   .0100235     0.21   0.831    -.0175089    .0217824
          5  |  -.0088071   .0105603    -0.83   0.404    -.0295048    .0118906
          6  |  -.0197509   .0144173    -1.37   0.171    -.0480083    .0085064
------------------------------------------------------------------------------

. marginsplot, level(95) ///
>         plotopts(msymbol(i) lcolor(black)) ///
>         ciopts(recast(rline) lpattern(dash) lcolor(black)) ///
>         title("") ///
>         xtitle("Individual education level", size(large) margin(medsmall)) ///
>         ytitle("Marginal effect of trade exposure", size(vlarge) margin(small)) ///
>         ylabel(, angle(hori) labsize(medsmall) nogrid) ///
>         xscale(noline range(-.5 5.5)) yscale(noline) ///
>         xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                 2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                 4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "', ///
>                         labsize(medsmall) nogrid) ///
>         yline(0, lcolor(black) lwidth(.2)) ///
>         graphregion(fcolor(white) ilcolor(white) lcolor(white) color(white) ///
>                 ifcolor(white) style(none)) ///
>         name(ladder_trade, replace) ///
>         addplot(hist educlvl, discrete yaxis(2) legend(off) ///
>                 ytitle("", axis(2)) ylabel(none, axis(2)) ///
>                 xlabel(0 `" "Primary" "or less" "' 1 `" "Primary" "completed" "' ///
>                         2 `" "Lower" "secondary" "' 3 `" "Higher" "secondary" "' ///
>                         4 `" "Secondary" "vocational" "' 5 `" "Tertiary" "completed" "',
>  ///
>                                 labsize(medsmall) nogrid) ///
>                 fcolor(none) lcolor(gs12))

  Variables that uniquely identify margins: educlvl

. 
. *** Figure 4 in Main Text 
. graph combine unemp_trade necess_trade satecon_trade ladder_trade, xcommon

. graph save "figure_individual", replace
(note: file figure_individual.gph not found)
(file figure_individual.gph saved)

. 
. *** Folder maintenance
. erase "glob_stats.dta"

. erase "map_coord.dta"

. erase "map_data.dta"

. erase "map_stats.dta"

. 
. 
. 
. ***** 7. Mechanism: individual grievances
. *------------------------------------------------------------------------------*
. 
. *** World Values Survey (Russia)
. * Load data
. use "PRW_2020_ISQ_WVSdata.dta", clear

. 
. *** Protest participation
. gen protestrec = 0 if E221B==2
(2,628 missing values generated)

. replace protestrec = 1 if E221B==1
(389 real changes made)

. 
. *** Grievance indicators
. * Life satisfaction
. gen lifesat = A170 if A170>0
(52 missing values generated)

. replace lifesat = (lifesat*(-1))+11
(4,481 real changes made)

. * Financial satisfaction
. gen econsat = C006 if C006>0
(55 missing values generated)

. replace econsat = (econsat*(-1))+11
(4,478 real changes made)

. * Income inequality
. gen inequal = E035 if E035>0
(242 missing values generated)

. replace inequal = (inequal*(-1))+11
(4,291 real changes made)

. * Not enough food (wave 6 only)
. gen nofood = H008_01 if H008_01>0
(2,051 missing values generated)

. recode nofood (1=4) (2=3) (3=2) (4=1)
(nofood: 2482 changes made)

. * No cash income (wave 6 only)
. gen nocash = H008_04 if H008_04>0
(2,059 missing values generated)

. recode nocash (1=4) (2=3) (3=2) (4=1)
(nocash: 2474 changes made)

. 
. *** Controls
. * Education
. gen educlvl = X025 if X025>0
(37 missing values generated)

. recode educlvl (1=1) (2=2) (3=3) (5=3) (4=4) (6=4) (7=5) (8=6)
(educlvl: 2373 changes made)

. * Female
. gen female = 0 if X001==1
(2,472 missing values generated)

. replace female = 1 if X001==2
(2,472 real changes made)

. * Age
. gen age = X003

. * Married
. gen married = 0 if X007==3 | X007==4 | X007==5 | X007==6
(2,631 missing values generated)

. replace married = 1 if X007==1 | X007==2
(2,589 real changes made)

. * Employment status
. gen unemp = X028==7

. * Income deciles
. gen income = X047 if X047>0
(411 missing values generated)

. * Politics important
. gen imppol = A004 if A004>0
(120 missing values generated)

. recode imppol (1=4) (2=3) (3=2) (4=1)
(imppol: 4413 changes made)

. * Federal district
. gen region = X048WVS

. 
. *** Table A3 in the Online Appendix
. tabstat protestrec lifesat econsat inequal nofood nocash female age income ///
>         married unemp imppol educlvl, statistics(N mean sd min max) format(%9.2f)

   stats |  protes~c   lifesat   econsat   inequal    nofood    nocash    female       age
---------+--------------------------------------------------------------------------------
       N |   2294.00   4481.00   4478.00   4291.00   2482.00   2474.00   4533.00   4533.00
    mean |      0.17      4.86      6.21      6.25      1.58      2.42      0.55     43.90
      sd |      0.38      2.30      2.47      3.34      0.84      1.03      0.50     17.19
     min |      0.00      1.00      1.00      1.00      1.00      1.00      0.00     16.00
     max |      1.00     10.00     10.00     10.00      4.00      4.00      1.00     91.00
------------------------------------------------------------------------------------------

   stats |    income   married     unemp    imppol   educlvl
---------+--------------------------------------------------
       N |   4122.00   4491.00   4533.00   4413.00   4496.00
    mean |      4.94      0.58      0.04      2.11      4.39
      sd |      2.20      0.49      0.20      0.89      1.06
     min |      1.00      0.00      0.00      1.00      1.00
     max |     10.00      1.00      1.00      4.00      6.00
------------------------------------------------------------

. 
. *** Regression models
. * Specification
. global C female age income married unemp imppol ib6.educlvl wave6

. * Models
. xtprobit protestrec lifesat $C, i(region) vce(cluster region)
note: 1.educlvl != 0 predicts failure perfectly
      1.educlvl dropped and 12 obs not used


Fitting comparison model:

Iteration 0:   log pseudolikelihood = -895.66735  
Iteration 1:   log pseudolikelihood = -808.15474  
Iteration 2:   log pseudolikelihood = -807.38248  
Iteration 3:   log pseudolikelihood = -807.38177  
Iteration 4:   log pseudolikelihood = -807.38177  

Fitting full model:

rho =  0.0     log pseudolikelihood = -807.38177
rho =  0.1     log pseudolikelihood = -816.26219

Iteration 0:   log pseudolikelihood = -814.91722  
Iteration 1:   log pseudolikelihood = -809.87736  (not concave)
Iteration 2:   log pseudolikelihood = -807.44436  (not concave)
Iteration 3:   log pseudolikelihood =  -807.4361  (not concave)
Iteration 4:   log pseudolikelihood = -807.43244  (not concave)
Iteration 5:   log pseudolikelihood = -807.42504  (not concave)
Iteration 6:   log pseudolikelihood = -807.42392  (not concave)
Iteration 7:   log pseudolikelihood = -807.42317  (not concave)
Iteration 8:   log pseudolikelihood =  -807.4225  (not concave)
Iteration 9:   log pseudolikelihood = -807.42174  (not concave)
Iteration 10:  log pseudolikelihood = -807.42061  (not concave)
Iteration 11:  log pseudolikelihood = -807.42004  (not concave)
Iteration 12:  log pseudolikelihood = -807.41956  (not concave)
Iteration 13:  log pseudolikelihood = -807.41911  (not concave)
Iteration 14:  log pseudolikelihood = -807.41865  (not concave)
Iteration 15:  log pseudolikelihood =  -807.4182  (not concave)
Iteration 16:  log pseudolikelihood = -807.41777  (not concave)
Iteration 17:  log pseudolikelihood = -807.41735  (not concave)
Iteration 18:  log pseudolikelihood = -807.41693  (not concave)
Iteration 19:  log pseudolikelihood = -807.41652  (not concave)
Iteration 20:  log pseudolikelihood = -807.41612  (not concave)
Iteration 21:  log pseudolikelihood = -807.41573  (not concave)
Iteration 22:  log pseudolikelihood = -807.41534  (not concave)
Iteration 23:  log pseudolikelihood = -807.41496  (not concave)
Iteration 24:  log pseudolikelihood = -807.41458  (not concave)
Iteration 25:  log pseudolikelihood = -807.41421  (not concave)
Iteration 26:  log pseudolikelihood = -807.41385  (not concave)
Iteration 27:  log pseudolikelihood = -807.41349  (not concave)
Iteration 28:  log pseudolikelihood = -807.41313  (not concave)
Iteration 29:  log pseudolikelihood = -807.41278  (not concave)
Iteration 30:  log pseudolikelihood = -807.41244  (not concave)
Iteration 31:  log pseudolikelihood =  -807.4121  (not concave)
Iteration 32:  log pseudolikelihood = -807.41176  (not concave)
Iteration 33:  log pseudolikelihood = -807.41143  (not concave)
Iteration 34:  log pseudolikelihood =  -807.4111  (not concave)
Iteration 35:  log pseudolikelihood = -807.41078  (not concave)
Iteration 36:  log pseudolikelihood = -807.41046  (not concave)
Iteration 37:  log pseudolikelihood = -807.41014  (not concave)
Iteration 38:  log pseudolikelihood = -807.40983  (not concave)
Iteration 39:  log pseudolikelihood = -807.40952  (not concave)
Iteration 40:  log pseudolikelihood = -807.40922  (not concave)
Iteration 41:  log pseudolikelihood = -807.40891  (not concave)
Iteration 42:  log pseudolikelihood = -807.40862  (not concave)
Iteration 43:  log pseudolikelihood = -807.40832  (not concave)
Iteration 44:  log pseudolikelihood = -807.40803  (not concave)
Iteration 45:  log pseudolikelihood = -807.40774  (not concave)
Iteration 46:  log pseudolikelihood = -807.40746  (not concave)
Iteration 47:  log pseudolikelihood = -807.40718  (not concave)
Iteration 48:  log pseudolikelihood =  -807.4069  (not concave)
Iteration 49:  log pseudolikelihood = -807.40663  (not concave)
Iteration 50:  log pseudolikelihood = -807.40635  (not concave)
Iteration 51:  log pseudolikelihood = -807.40608  (not concave)
Iteration 52:  log pseudolikelihood = -807.40582  (not concave)
Iteration 53:  log pseudolikelihood = -807.40556  (not concave)
Iteration 54:  log pseudolikelihood =  -807.4053  (not concave)
Iteration 55:  log pseudolikelihood = -807.40504  (not concave)
Iteration 56:  log pseudolikelihood = -807.40479  (not concave)
Iteration 57:  log pseudolikelihood = -807.40453  (not concave)
Iteration 58:  log pseudolikelihood = -807.40429  (not concave)
Iteration 59:  log pseudolikelihood = -807.40404  (not concave)
Iteration 60:  log pseudolikelihood =  -807.4038  (not concave)
Iteration 61:  log pseudolikelihood = -807.40356  (not concave)
Iteration 62:  log pseudolikelihood = -807.40332  (not concave)
Iteration 63:  log pseudolikelihood = -807.40308  (not concave)
Iteration 64:  log pseudolikelihood = -807.40285  (not concave)
Iteration 65:  log pseudolikelihood = -807.40262  (not concave)
Iteration 66:  log pseudolikelihood = -807.40239  (not concave)
Iteration 67:  log pseudolikelihood = -807.40217  (not concave)
Iteration 68:  log pseudolikelihood = -807.40195  (not concave)
Iteration 69:  log pseudolikelihood = -807.40173  (not concave)
Iteration 70:  log pseudolikelihood = -807.40151  (not concave)
Iteration 71:  log pseudolikelihood = -807.40129  (not concave)
Iteration 72:  log pseudolikelihood = -807.40108  (not concave)
Iteration 73:  log pseudolikelihood = -807.40087  (not concave)
Iteration 74:  log pseudolikelihood = -807.40066  (not concave)
Iteration 75:  log pseudolikelihood = -807.40045  (not concave)
Iteration 76:  log pseudolikelihood = -807.40025  (not concave)
Iteration 77:  log pseudolikelihood = -807.40005  
Iteration 78:  log pseudolikelihood = -807.38177  
Iteration 79:  log pseudolikelihood = -807.38177  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,849
Group variable: region                          Number of groups  =          8

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         54
                                                              avg =      231.1
                                                              max =        545

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =          .
Log pseudolikelihood  = -807.38177              Prob > chi2       =          .

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     lifesat |   .0355269   .0156876     2.26   0.024     .0047797    .0662741
      female |  -.0603069   .0664913    -0.91   0.364    -.1906274    .0700136
         age |   .0058196   .0022689     2.56   0.010     .0013727    .0102666
      income |   .0161066   .0189359     0.85   0.395     -.021007    .0532202
     married |   .0713404    .056638     1.26   0.208     -.039668    .1823488
       unemp |  -.3241504   .1237924    -2.62   0.009    -.5667789   -.0815218
      imppol |   .1265132   .0398926     3.17   0.002     .0483252    .2047011
             |
     educlvl |
          1  |          0  (empty)
          2  |   .0821797   .2235169     0.37   0.713    -.3559053    .5202647
          3  |  -.4126602   .1076755    -3.83   0.000    -.6237004     -.20162
          4  |  -.1806024   .1403691    -1.29   0.198    -.4557208    .0945161
          5  |  -.2602413   .2796415    -0.93   0.352    -.8083285    .2878459
             |
       wave6 |   .9726407   .2755958     3.53   0.000     .4324829    1.512799
       _cons |  -1.732947   .2103371    -8.24   0.000      -2.1452   -1.320694
-------------+----------------------------------------------------------------
    /lnsig2u |  -38.81359          .                             .           .
-------------+----------------------------------------------------------------
     sigma_u |   3.73e-09          .                             .           .
         rho |   1.39e-17          .                             .           .
------------------------------------------------------------------------------

. est store prot1

. xtprobit protestrec econsat $C, i(region) vce(cluster region)
note: 1.educlvl != 0 predicts failure perfectly
      1.educlvl dropped and 12 obs not used


Fitting comparison model:

Iteration 0:   log pseudolikelihood = -891.70317  
Iteration 1:   log pseudolikelihood = -805.95449  
Iteration 2:   log pseudolikelihood = -805.24933  
Iteration 3:   log pseudolikelihood = -805.24873  
Iteration 4:   log pseudolikelihood = -805.24873  

Fitting full model:

rho =  0.0     log pseudolikelihood = -805.24873
rho =  0.1     log pseudolikelihood = -813.90433

Iteration 0:   log pseudolikelihood = -812.65611  
Iteration 1:   log pseudolikelihood = -807.44017  (not concave)
Iteration 2:   log pseudolikelihood = -805.47966  (not concave)
Iteration 3:   log pseudolikelihood = -805.29399  (not concave)
Iteration 4:   log pseudolikelihood = -805.29241  
Iteration 5:   log pseudolikelihood = -805.24873  
Iteration 6:   log pseudolikelihood = -805.24873  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,844
Group variable: region                          Number of groups  =          8

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         54
                                                              avg =      230.5
                                                              max =        546

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =          .
Log pseudolikelihood  = -805.24873              Prob > chi2       =          .

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     econsat |   .0192327   .0109761     1.75   0.080      -.00228    .0407455
      female |  -.0779997   .0652042    -1.20   0.232    -.2057975    .0497982
         age |   .0066692   .0023023     2.90   0.004     .0021568    .0111815
      income |   .0148479   .0216315     0.69   0.492     -.027549    .0572448
     married |    .063847   .0591738     1.08   0.281    -.0521315    .1798255
       unemp |  -.3078396     .12539    -2.46   0.014    -.5535995   -.0620796
      imppol |   .1225345   .0387522     3.16   0.002     .0465816    .1984874
             |
     educlvl |
          1  |          0  (empty)
          2  |   .0023893   .2208168     0.01   0.991    -.4304037    .4351823
          3  |  -.4296015   .1202997    -3.57   0.000    -.6653845   -.1938186
          4  |  -.1895906    .138993    -1.36   0.173    -.4620119    .0828307
          5  |  -.2410002   .2800885    -0.86   0.390    -.7899637    .3079632
             |
       wave6 |    .968888   .2846473     3.40   0.001     .4109895    1.526787
       _cons |   -1.68103   .2080511    -8.08   0.000    -2.088802   -1.273257
-------------+----------------------------------------------------------------
    /lnsig2u |  -44.78385    2096927                      -4109945     4109856
-------------+----------------------------------------------------------------
     sigma_u |   1.88e-10   .0001976                             0           .
         rho |   3.55e-20   7.45e-14                             0           .
------------------------------------------------------------------------------

. est store prot2

. xtprobit protestrec inequal $C, i(region) vce(cluster region)
note: 1.educlvl != 0 predicts failure perfectly
      1.educlvl dropped and 11 obs not used


Fitting comparison model:

Iteration 0:   log pseudolikelihood = -877.50928  
Iteration 1:   log pseudolikelihood = -796.02957  
Iteration 2:   log pseudolikelihood = -795.40138  
Iteration 3:   log pseudolikelihood =  -795.4009  
Iteration 4:   log pseudolikelihood =  -795.4009  

Fitting full model:

rho =  0.0     log pseudolikelihood =  -795.4009
rho =  0.1     log pseudolikelihood = -804.29811

Iteration 0:   log pseudolikelihood = -802.89308  
Iteration 1:   log pseudolikelihood = -796.73538  
Iteration 2:   log pseudolikelihood = -795.54623  
Iteration 3:   log pseudolikelihood = -795.40091  
Iteration 4:   log pseudolikelihood =  -795.4009  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,818
Group variable: region                          Number of groups  =          8

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         50
                                                              avg =      227.2
                                                              max =        539

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =          .
Log pseudolikelihood  =  -795.4009              Prob > chi2       =          .

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     inequal |   .0094996   .0052846     1.80   0.072    -.0008581    .0198572
      female |  -.0740829   .0648433    -1.14   0.253    -.2011734    .0530076
         age |   .0060733    .002306     2.63   0.008     .0015537    .0105928
      income |   .0107102   .0216461     0.49   0.621    -.0317154    .0531357
     married |    .051416   .0600283     0.86   0.392    -.0662373    .1690693
       unemp |  -.2929257   .1260371    -2.32   0.020    -.5399539   -.0458975
      imppol |    .121941   .0389303     3.13   0.002      .045639    .1982431
             |
     educlvl |
          1  |          0  (empty)
          2  |  -.0081328   .2759969    -0.03   0.976    -.5490767    .5328111
          3  |  -.4155704   .1027949    -4.04   0.000    -.6170448    -.214096
          4  |   -.187897     .14529    -1.29   0.196    -.4726602    .0968661
          5  |  -.2638703   .2756716    -0.96   0.338    -.8041767    .2764361
             |
       wave6 |   .9218363    .271171     3.40   0.001     .3903509    1.453322
       _cons |  -1.543284   .2414844    -6.39   0.000    -2.016585   -1.069983
-------------+----------------------------------------------------------------
    /lnsig2u |  -42.72211          .                             .           .
-------------+----------------------------------------------------------------
     sigma_u |   5.28e-10          .                             .           .
         rho |   2.79e-19          .                             .           .
------------------------------------------------------------------------------

. est store prot3

. xtprobit protestrec nofood $C, i(region) vce(cluster region)
note: wave6 omitted because of collinearity

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -188.25917  
Iteration 1:   log pseudolikelihood = -176.46706  
Iteration 2:   log pseudolikelihood = -176.43735  
Iteration 3:   log pseudolikelihood = -176.43733  
Iteration 4:   log pseudolikelihood = -176.43733  

Fitting full model:

rho =  0.0     log pseudolikelihood = -176.43733
rho =  0.1     log pseudolikelihood = -172.03037
rho =  0.2     log pseudolikelihood = -173.26288

Iteration 0:   log pseudolikelihood = -172.00523  
Iteration 1:   log pseudolikelihood =  -171.5257  
Iteration 2:   log pseudolikelihood = -171.52491  
Iteration 3:   log pseudolikelihood = -171.52491  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =        273
Group variable: region                          Number of groups  =          8

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          8
                                                              avg =       34.1
                                                              max =         71

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =          .
Log pseudolikelihood  = -171.52491              Prob > chi2       =          .

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      nofood |   .1199651   .0445428     2.69   0.007     .0326627    .2072674
      female |  -.0320577   .1041677    -0.31   0.758    -.2362227    .1721073
         age |   -.004261   .0085838    -0.50   0.620     -.021085     .012563
      income |   .0395254   .0644483     0.61   0.540     -.086791    .1658418
     married |   .2198199   .0996616     2.21   0.027     .0244868     .415153
       unemp |   1.011888   .7318256     1.38   0.167    -.4224637     2.44624
      imppol |   .2170965   .0600182     3.62   0.000      .099463      .33473
             |
     educlvl |
          2  |  -.4375092   .9645795    -0.45   0.650     -2.32805    1.453032
          3  |  -.5190609   .2846917    -1.82   0.068    -1.077046    .0389246
          4  |   .0625151   .1693709     0.37   0.712    -.2694458    .3944759
          5  |  -.4591368   .4293634    -1.07   0.285    -1.300674    .3824001
             |
       wave6 |          0  (omitted)
       _cons |  -.7782343   .5536978    -1.41   0.160    -1.863462    .3069935
-------------+----------------------------------------------------------------
    /lnsig2u |   -2.40677   .8051996                     -3.984932   -.8286076
-------------+----------------------------------------------------------------
     sigma_u |   .3001764    .120851                      .1363587    .6608002
         rho |   .0826579   .0610547                      .0182543    .3039396
------------------------------------------------------------------------------

. est store prot4

. xtprobit protestrec nocash $C, i(region) vce(cluster region)
note: wave6 omitted because of collinearity

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -187.02842  
Iteration 1:   log pseudolikelihood = -173.94374  
Iteration 2:   log pseudolikelihood = -173.90497  
Iteration 3:   log pseudolikelihood = -173.90495  
Iteration 4:   log pseudolikelihood = -173.90495  

Fitting full model:

rho =  0.0     log pseudolikelihood = -173.90495
rho =  0.1     log pseudolikelihood = -170.23329
rho =  0.2     log pseudolikelihood = -171.51716

Iteration 0:   log pseudolikelihood = -170.21217  
Iteration 1:   log pseudolikelihood = -169.73001  
Iteration 2:   log pseudolikelihood = -169.72906  
Iteration 3:   log pseudolikelihood = -169.72906  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =        271
Group variable: region                          Number of groups  =          8

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          8
                                                              avg =       33.9
                                                              max =         71

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =          .
Log pseudolikelihood  = -169.72906              Prob > chi2       =          .

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      nocash |   .1891484   .0608763     3.11   0.002     .0698331    .3084636
      female |  -.0264011   .0994445    -0.27   0.791    -.2213087    .1685066
         age |  -.0022374   .0086117    -0.26   0.795     -.019116    .0146411
      income |   .0742099   .0617734     1.20   0.230    -.0468638    .1952835
     married |   .2237627   .1123373     1.99   0.046     .0035857    .4439398
       unemp |   1.064085   .7229687     1.47   0.141    -.3529076    2.481077
      imppol |   .2070194   .0647426     3.20   0.001     .0801262    .3339127
             |
     educlvl |
          2  |  -.2936654   .9883484    -0.30   0.766    -2.230793    1.643462
          3  |  -.4731144   .2651184    -1.78   0.074    -.9927369    .0465081
          4  |   .1281765   .1623306     0.79   0.430    -.1899856    .4463385
          5  |  -.2846003   .4003742    -0.71   0.477    -1.069319    .5001187
             |
       wave6 |          0  (omitted)
       _cons |  -1.340198    .446776    -3.00   0.003    -2.215863   -.4645329
-------------+----------------------------------------------------------------
    /lnsig2u |  -2.516017   .8160988                     -4.115541   -.9164926
-------------+----------------------------------------------------------------
     sigma_u |   .2842195   .1159756                      .1277384    .6323917
         rho |   .0747429   .0564385                      .0160551    .2856731
------------------------------------------------------------------------------

. est store prot5

. * Table 4 in Main Text
. esttab prot1 prot2 prot3 prot4 prot5, ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g r2_p p, labels("# of observations" "# of regions" ///
>                 "Pseudo R squared " "Prob > Chi2") fmt(0 0 2 3)) ///
>         order(lifesat econsat inequal nofood nocash)

------------------------------------------------------------------------------------------
> ----------
                              (1)             (2)             (3)             (4)         
>     (5)   
                       protestrec      protestrec      protestrec      protestrec      pro
> testrec   
                             b/se            b/se            b/se            b/se         
>    b/se   
------------------------------------------------------------------------------------------
> ----------
protestrec                                                                                
>           
lifesat                     0.036**                                                       
>           
                           (0.02)                                                         
>           
econsat                                     0.019*                                        
>           
                                           (0.01)                                         
>           
inequal                                                     0.009*                        
>           
                                                           (0.01)                         
>           
nofood                                                                      0.120***      
>           
                                                                           (0.04)         
>           
nocash                                                                                    
>   0.189***
                                                                                          
>  (0.06)   
female                     -0.060          -0.078          -0.074          -0.032         
>  -0.026   
                           (0.07)          (0.07)          (0.06)          (0.10)         
>  (0.10)   
age                         0.006**         0.007***        0.006***       -0.004         
>  -0.002   
                           (0.00)          (0.00)          (0.00)          (0.01)         
>  (0.01)   
income                      0.016           0.015           0.011           0.040         
>   0.074   
                           (0.02)          (0.02)          (0.02)          (0.06)         
>  (0.06)   
married                     0.071           0.064           0.051           0.220**       
>   0.224** 
                           (0.06)          (0.06)          (0.06)          (0.10)         
>  (0.11)   
unemp                      -0.324***       -0.308**        -0.293**         1.012         
>   1.064   
                           (0.12)          (0.13)          (0.13)          (0.73)         
>  (0.72)   
imppol                      0.127***        0.123***        0.122***        0.217***      
>   0.207***
                           (0.04)          (0.04)          (0.04)          (0.06)         
>  (0.06)   
educlvl=1                   0.000           0.000           0.000                         
>           
                              (.)             (.)             (.)                         
>           
educlvl=2                   0.082           0.002          -0.008          -0.438         
>  -0.294   
                           (0.22)          (0.22)          (0.28)          (0.96)         
>  (0.99)   
educlvl=3                  -0.413***       -0.430***       -0.416***       -0.519*        
>  -0.473*  
                           (0.11)          (0.12)          (0.10)          (0.28)         
>  (0.27)   
educlvl=4                  -0.181          -0.190          -0.188           0.063         
>   0.128   
                           (0.14)          (0.14)          (0.15)          (0.17)         
>  (0.16)   
educlvl=5                  -0.260          -0.241          -0.264          -0.459         
>  -0.285   
                           (0.28)          (0.28)          (0.28)          (0.43)         
>  (0.40)   
educlvl=6                   0.000           0.000           0.000           0.000         
>   0.000   
                              (.)             (.)             (.)             (.)         
>     (.)   
wave6                       0.973***        0.969***        0.922***        0.000         
>   0.000   
                           (0.28)          (0.28)          (0.27)             (.)         
>     (.)   
Constant                   -1.733***       -1.681***       -1.543***       -0.778         
>  -1.340***
                           (0.21)          (0.21)          (0.24)          (0.55)         
>  (0.45)   
------------------------------------------------------------------------------------------
> ----------
lnsig2u                                                                                   
>           
Constant                  -38.814         -44.784         -42.722          -2.407***      
>  -2.516***
                              (.)    (2096926.64)             (.)          (0.81)         
>  (0.82)   
------------------------------------------------------------------------------------------
> ----------
# of observations            1849            1844            1818             273         
>     271   
# of regions                    8               8               8               8         
>       8   
Pseudo R squared                                                                          
>           
Prob > Chi2                     .               .               .               .         
>       .   
------------------------------------------------------------------------------------------
> ----------

. 
. *** Regression models
. * Specification
. global C female age income married unemp imppol ib6.educlvl wave6 i.region

. * Models
. probit protestrec lifesat $C, cluster(region)

note: 1.educlvl != 0 predicts failure perfectly
      1.educlvl dropped and 12 obs not used

Iteration 0:   log pseudolikelihood = -895.66735  
Iteration 1:   log pseudolikelihood = -806.05109  
Iteration 2:   log pseudolikelihood =   -805.211  
Iteration 3:   log pseudolikelihood =  -805.2102  
Iteration 4:   log pseudolikelihood =  -805.2102  

Probit regression                               Number of obs     =      1,849
                                                Wald chi2(6)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood =  -805.2102               Pseudo R2         =     0.1010

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     lifesat |   .0370251   .0153056     2.42   0.016     .0070267    .0670236
      female |  -.0623263   .0625861    -1.00   0.319    -.1849929    .0603402
         age |   .0061158   .0021857     2.80   0.005      .001832    .0103996
      income |   .0237738   .0198578     1.20   0.231    -.0151469    .0626944
     married |   .0639352   .0606729     1.05   0.292    -.0549814    .1828519
       unemp |  -.3374991   .1238837    -2.72   0.006    -.5803066   -.0946916
      imppol |   .1232062   .0396611     3.11   0.002     .0454718    .2009406
             |
     educlvl |
          1  |          0  (empty)
          2  |   .0852543   .2262494     0.38   0.706    -.3581864    .5286949
          3  |   -.429118   .1085286    -3.95   0.000    -.6418301   -.2164059
          4  |  -.1896201   .1444792    -1.31   0.189    -.4727941    .0935539
          5  |  -.2477488   .2826929    -0.88   0.381    -.8018168    .3063191
             |
       wave6 |   .9934944   .3240654     3.07   0.002     .3583379    1.628651
             |
      region |
     643003  |  -.0651223   .0392701    -1.66   0.097    -.1420903    .0118458
     643007  |   .0727316   .0118127     6.16   0.000     .0495791    .0958841
     643008  |   .0403636   .0185962     2.17   0.030     .0039157    .0768115
     643011  |  -.1689289   .0331917    -5.09   0.000    -.2339835   -.1038743
     643012  |  -.0625326   .2419884    -0.26   0.796    -.5368211    .4117559
     643013  |  -.0648842   .0123059    -5.27   0.000    -.0890034   -.0407651
     643014  |   -.179318    .030766    -5.83   0.000    -.2396181   -.1190178
             |
       _cons |  -1.743106   .2073233    -8.41   0.000    -2.149452   -1.336759
------------------------------------------------------------------------------

. est store prot6

. probit protestrec econsat $C, cluster(region)

note: 1.educlvl != 0 predicts failure perfectly
      1.educlvl dropped and 12 obs not used

Iteration 0:   log pseudolikelihood = -891.70317  
Iteration 1:   log pseudolikelihood = -803.71078  
Iteration 2:   log pseudolikelihood = -802.93438  
Iteration 3:   log pseudolikelihood =  -802.9337  
Iteration 4:   log pseudolikelihood =  -802.9337  

Probit regression                               Number of obs     =      1,844
                                                Wald chi2(6)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood =  -802.9337               Pseudo R2         =     0.0996

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     econsat |    .020275   .0115118     1.76   0.078    -.0022878    .0428378
      female |  -.0803578   .0608292    -1.32   0.186    -.1995808    .0388652
         age |   .0070083   .0022263     3.15   0.002     .0026449    .0113717
      income |   .0231826   .0224842     1.03   0.303    -.0208857    .0672509
     married |   .0559386    .063138     0.89   0.376    -.0678095    .1796868
       unemp |    -.32213   .1260535    -2.56   0.011    -.5691902   -.0750698
      imppol |   .1188487   .0387455     3.07   0.002     .0429088    .1947885
             |
     educlvl |
          1  |          0  (empty)
          2  |    .002347   .2251695     0.01   0.992    -.4389771    .4436711
          3  |  -.4488159   .1229875    -3.65   0.000    -.6898669   -.2077649
          4  |   -.198982   .1433497    -1.39   0.165    -.4799421    .0819782
          5  |  -.2255126   .2829406    -0.80   0.425     -.780066    .3290408
             |
       wave6 |   .9891041   .3355686     2.95   0.003     .3314017    1.646807
             |
      region |
     643003  |  -.0517323   .0475132    -1.09   0.276    -.1448564    .0413919
     643007  |    .081158   .0103864     7.81   0.000     .0608011    .1015149
     643008  |    .053507   .0201811     2.65   0.008     .0139528    .0930612
     643011  |  -.1558222   .0371455    -4.19   0.000    -.2286261   -.0830183
     643012  |  -.0553408   .2553707    -0.22   0.828    -.5558581    .4451766
     643013  |  -.0653584   .0130805    -5.00   0.000    -.0909958   -.0397211
     643014  |  -.1865441   .0322069    -5.79   0.000    -.2496684   -.1234198
             |
       _cons |  -1.699404   .2132776    -7.97   0.000    -2.117421   -1.281388
------------------------------------------------------------------------------

. est store prot7

. probit protestrec inequal $C, cluster(region)

note: 1.educlvl != 0 predicts failure perfectly
      1.educlvl dropped and 11 obs not used

Iteration 0:   log pseudolikelihood = -877.50928  
Iteration 1:   log pseudolikelihood = -793.89104  
Iteration 2:   log pseudolikelihood =   -793.205  
Iteration 3:   log pseudolikelihood = -793.20445  
Iteration 4:   log pseudolikelihood = -793.20445  

Probit regression                               Number of obs     =      1,818
                                                Wald chi2(6)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood = -793.20445               Pseudo R2         =     0.0961

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     inequal |   .0097753   .0053012     1.84   0.065    -.0006149    .0201655
      female |  -.0766037   .0605865    -1.26   0.206    -.1953511    .0421436
         age |   .0064445   .0022374     2.88   0.004     .0020593    .0108297
      income |    .018835   .0218644     0.86   0.389    -.0240185    .0616884
     married |   .0436486   .0641198     0.68   0.496    -.0820238     .169321
       unemp |  -.3126612   .1245915    -2.51   0.012     -.556856   -.0684663
      imppol |   .1181625   .0389531     3.03   0.002     .0418158    .1945093
             |
     educlvl |
          1  |          0  (empty)
          2  |  -.0069517   .2745682    -0.03   0.980    -.5450955    .5311921
          3  |  -.4345415   .1053719    -4.12   0.000    -.6410667   -.2280164
          4  |  -.1948023   .1500455    -1.30   0.194    -.4888861    .0992816
          5  |  -.2487643   .2777452    -0.90   0.370    -.7931349    .2956063
             |
       wave6 |   .9171428   .3175765     2.89   0.004     .2947043    1.539581
             |
      region |
     643003  |  -.0046613   .0335205    -0.14   0.889    -.0703603    .0610376
     643007  |     .11832   .0214558     5.51   0.000     .0762675    .1603725
     643008  |   .0858325   .0287566     2.98   0.003     .0294706    .1421944
     643011  |   -.109135   .0343615    -3.18   0.001    -.1764822   -.0417878
     643012  |   .0938577   .2341129     0.40   0.688    -.3649952    .5527107
     643013  |  -.0162098     .01847    -0.88   0.380    -.0524103    .0199907
     643014  |  -.1511073   .0262052    -5.77   0.000    -.2024685    -.099746
             |
       _cons |  -1.596273   .2440377    -6.54   0.000    -2.074578   -1.117967
------------------------------------------------------------------------------

. est store prot8

. probit protestrec nofood $C, cluster(region)

note: wave6 omitted because of collinearity
Iteration 0:   log pseudolikelihood = -188.25917  
Iteration 1:   log pseudolikelihood = -164.65624  
Iteration 2:   log pseudolikelihood = -164.56704  
Iteration 3:   log pseudolikelihood = -164.56696  
Iteration 4:   log pseudolikelihood = -164.56696  

Probit regression                               Number of obs     =        273
                                                Wald chi2(6)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood = -164.56696               Pseudo R2         =     0.1258

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      nofood |   .1081818   .0440441     2.46   0.014     .0218569    .1945067
      female |  -.0368699   .1074172    -0.34   0.731    -.2474036    .1736639
         age |   -.004663   .0084242    -0.55   0.580    -.0211742    .0118482
      income |   .0396502   .0645139     0.61   0.539    -.0867948    .1660952
     married |    .170985   .0910731     1.88   0.060     -.007515     .349485
       unemp |   1.015243   .7419713     1.37   0.171    -.4389944     2.46948
      imppol |   .2096226   .0619922     3.38   0.001       .08812    .3311252
             |
     educlvl |
          2  |  -.4609838   .9731566    -0.47   0.636    -2.368336    1.446368
          3  |   -.526675   .2911636    -1.81   0.070    -1.097345    .0439951
          4  |   .0706969    .170367     0.41   0.678    -.2632164    .4046101
          5  |  -.5341354   .4197532    -1.27   0.203    -1.356837    .2885658
             |
       wave6 |          0  (omitted)
             |
      region |
     643003  |  -.1600089   .0831355    -1.92   0.054    -.3229514    .0029336
     643007  |    .228868   .0915393     2.50   0.012     .0494543    .4082817
     643008  |   .3367278   .0845288     3.98   0.000     .1710543    .5024012
     643011  |   .0090002   .1206203     0.07   0.941    -.2274113    .2454117
     643012  |   .2574386   .0824489     3.12   0.002     .0958418    .4190355
     643013  |   .9855133   .1020835     9.65   0.000     .7854333    1.185593
     643014  |   .1049092   .0874282     1.20   0.230    -.0664469    .2762653
             |
       _cons |  -.9400667   .5644757    -1.67   0.096    -2.046419    .1662853
------------------------------------------------------------------------------

. est store prot9

. probit protestrec nocash $C, cluster(region)

note: wave6 omitted because of collinearity
Iteration 0:   log pseudolikelihood = -187.02842  
Iteration 1:   log pseudolikelihood =  -163.1771  
Iteration 2:   log pseudolikelihood = -163.09099  
Iteration 3:   log pseudolikelihood = -163.09091  
Iteration 4:   log pseudolikelihood = -163.09091  

Probit regression                               Number of obs     =        271
                                                Wald chi2(6)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood = -163.09091               Pseudo R2         =     0.1280

                                 (Std. Err. adjusted for 8 clusters in region)
------------------------------------------------------------------------------
             |               Robust
  protestrec |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      nocash |   .1768836   .0607502     2.91   0.004     .0578154    .2959519
      female |  -.0320111   .1023449    -0.31   0.754    -.2326035    .1685814
         age |   -.002806   .0084834    -0.33   0.741    -.0194332    .0138212
      income |   .0730885   .0625758     1.17   0.243    -.0495579    .1957349
     married |   .1767997   .1068047     1.66   0.098    -.0325337     .386133
       unemp |   1.063193    .733272     1.45   0.147    -.3739939     2.50038
      imppol |   .1978015   .0666821     2.97   0.003     .0671069     .328496
             |
     educlvl |
          2  |  -.3190918   .9967121    -0.32   0.749    -2.272612    1.634428
          3  |  -.4845622   .2719867    -1.78   0.075    -1.017646    .0485219
          4  |   .1302112   .1626654     0.80   0.423    -.1886071    .4490295
          5  |  -.3779182   .3959256    -0.95   0.340    -1.153918    .3980817
             |
       wave6 |          0  (omitted)
             |
      region |
     643003  |  -.1890543   .0832723    -2.27   0.023     -.352265   -.0258436
     643007  |   .2249116   .0900952     2.50   0.013     .0483282     .401495
     643008  |   .2753445   .0819644     3.36   0.001     .1146971    .4359919
     643011  |  -.0196022   .1151907    -0.17   0.865    -.2453719    .2061675
     643012  |   .1743331   .0762723     2.29   0.022     .0248422     .323824
     643013  |   .9162656   .0951755     9.63   0.000      .729725    1.102806
     643014  |   .0930242    .086845     1.07   0.284    -.0771889    .2632372
             |
       _cons |   -1.42833    .457548    -3.12   0.002    -2.325108   -.5315527
------------------------------------------------------------------------------

. est store prot10

. * Table
. esttab prot6 prot7 prot8 prot9 prot10, ///
>         cells(b(star fmt(%9.3f)) se(par fmt(2))) style(fixed) ///
>         starlevels(* 0.10 ** 0.05 *** 0.01) label ///
>         stats(N N_g r2_p p, labels("# of individuals" "# of regions" ///
>                 "Pseudo R squared " "Prob > Chi2") fmt(0 0 2 3)) ///
>         order(lifesat econsat inequal nofood nocash)

------------------------------------------------------------------------------------------
> ----------
                              (1)             (2)             (3)             (4)         
>     (5)   
                       protestrec      protestrec      protestrec      protestrec      pro
> testrec   
                             b/se            b/se            b/se            b/se         
>    b/se   
------------------------------------------------------------------------------------------
> ----------
protestrec                                                                                
>           
lifesat                     0.037**                                                       
>           
                           (0.02)                                                         
>           
econsat                                     0.020*                                        
>           
                                           (0.01)                                         
>           
inequal                                                     0.010*                        
>           
                                                           (0.01)                         
>           
nofood                                                                      0.108**       
>           
                                                                           (0.04)         
>           
nocash                                                                                    
>   0.177***
                                                                                          
>  (0.06)   
female                     -0.062          -0.080          -0.077          -0.037         
>  -0.032   
                           (0.06)          (0.06)          (0.06)          (0.11)         
>  (0.10)   
age                         0.006***        0.007***        0.006***       -0.005         
>  -0.003   
                           (0.00)          (0.00)          (0.00)          (0.01)         
>  (0.01)   
income                      0.024           0.023           0.019           0.040         
>   0.073   
                           (0.02)          (0.02)          (0.02)          (0.06)         
>  (0.06)   
married                     0.064           0.056           0.044           0.171*        
>   0.177*  
                           (0.06)          (0.06)          (0.06)          (0.09)         
>  (0.11)   
unemp                      -0.337***       -0.322**        -0.313**         1.015         
>   1.063   
                           (0.12)          (0.13)          (0.12)          (0.74)         
>  (0.73)   
imppol                      0.123***        0.119***        0.118***        0.210***      
>   0.198***
                           (0.04)          (0.04)          (0.04)          (0.06)         
>  (0.07)   
educlvl=1                   0.000           0.000           0.000                         
>           
                              (.)             (.)             (.)                         
>           
educlvl=2                   0.085           0.002          -0.007          -0.461         
>  -0.319   
                           (0.23)          (0.23)          (0.27)          (0.97)         
>  (1.00)   
educlvl=3                  -0.429***       -0.449***       -0.435***       -0.527*        
>  -0.485*  
                           (0.11)          (0.12)          (0.11)          (0.29)         
>  (0.27)   
educlvl=4                  -0.190          -0.199          -0.195           0.071         
>   0.130   
                           (0.14)          (0.14)          (0.15)          (0.17)         
>  (0.16)   
educlvl=5                  -0.248          -0.226          -0.249          -0.534         
>  -0.378   
                           (0.28)          (0.28)          (0.28)          (0.42)         
>  (0.40)   
educlvl=6                   0.000           0.000           0.000           0.000         
>   0.000   
                              (.)             (.)             (.)             (.)         
>     (.)   
wave6                       0.993***        0.989***        0.917***        0.000         
>   0.000   
                           (0.32)          (0.34)          (0.32)             (.)         
>     (.)   
region=643002               0.000           0.000           0.000           0.000         
>   0.000   
                              (.)             (.)             (.)             (.)         
>     (.)   
region=643003              -0.065*         -0.052          -0.005          -0.160*        
>  -0.189** 
                           (0.04)          (0.05)          (0.03)          (0.08)         
>  (0.08)   
region=643007               0.073***        0.081***        0.118***        0.229**       
>   0.225** 
                           (0.01)          (0.01)          (0.02)          (0.09)         
>  (0.09)   
region=643008               0.040**         0.054***        0.086***        0.337***      
>   0.275***
                           (0.02)          (0.02)          (0.03)          (0.08)         
>  (0.08)   
region=643011              -0.169***       -0.156***       -0.109***        0.009         
>  -0.020   
                           (0.03)          (0.04)          (0.03)          (0.12)         
>  (0.12)   
region=643012              -0.063          -0.055           0.094           0.257***      
>   0.174** 
                           (0.24)          (0.26)          (0.23)          (0.08)         
>  (0.08)   
region=643013              -0.065***       -0.065***       -0.016           0.986***      
>   0.916***
                           (0.01)          (0.01)          (0.02)          (0.10)         
>  (0.10)   
region=643014              -0.179***       -0.187***       -0.151***        0.105         
>   0.093   
                           (0.03)          (0.03)          (0.03)          (0.09)         
>  (0.09)   
Constant                   -1.743***       -1.699***       -1.596***       -0.940*        
>  -1.428***
                           (0.21)          (0.21)          (0.24)          (0.56)         
>  (0.46)   
------------------------------------------------------------------------------------------
> ----------
# of individuals             1849            1844            1818             273         
>     271   
# of regions                                                                              
>           
Pseudo R squared             0.10            0.10            0.10            0.13         
>    0.13   
Prob > Chi2                     .               .               .               .         
>       .   
------------------------------------------------------------------------------------------
> ----------

. 
. 
. 
. *------------------------------------------------------------------------------*
. 
. 
. capture log close
